{
  "best_metric": 3.3018171787261963,
  "best_model_checkpoint": "/scratch/cl5625/exceptions/models/100M_495/checkpoint-90000",
  "epoch": 10.0,
  "eval_steps": 1000,
  "global_step": 92910,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.005381552039608223,
      "grad_norm": 2.1159660816192627,
      "learning_rate": 0.00028199999999999997,
      "loss": 8.9316,
      "step": 50
    },
    {
      "epoch": 0.010763104079216447,
      "grad_norm": 1.9895623922348022,
      "learning_rate": 0.0005819999999999999,
      "loss": 6.9096,
      "step": 100
    },
    {
      "epoch": 0.01614465611882467,
      "grad_norm": 1.8703136444091797,
      "learning_rate": 0.0005996961534317422,
      "loss": 6.4463,
      "step": 150
    },
    {
      "epoch": 0.021526208158432893,
      "grad_norm": 1.2748037576675415,
      "learning_rate": 0.0005993729124016807,
      "loss": 6.2017,
      "step": 200
    },
    {
      "epoch": 0.026907760198041114,
      "grad_norm": 1.3479403257369995,
      "learning_rate": 0.0005990496713716194,
      "loss": 6.1031,
      "step": 250
    },
    {
      "epoch": 0.03228931223764934,
      "grad_norm": 1.0388602018356323,
      "learning_rate": 0.000598726430341558,
      "loss": 5.9726,
      "step": 300
    },
    {
      "epoch": 0.03767086427725756,
      "grad_norm": 1.1332666873931885,
      "learning_rate": 0.0005984031893114966,
      "loss": 5.886,
      "step": 350
    },
    {
      "epoch": 0.04305241631686579,
      "grad_norm": 1.398230791091919,
      "learning_rate": 0.0005980799482814351,
      "loss": 5.8019,
      "step": 400
    },
    {
      "epoch": 0.048433968356474004,
      "grad_norm": 1.5357556343078613,
      "learning_rate": 0.0005977567072513737,
      "loss": 5.7437,
      "step": 450
    },
    {
      "epoch": 0.05381552039608223,
      "grad_norm": 1.175200343132019,
      "learning_rate": 0.0005974334662213123,
      "loss": 5.6616,
      "step": 500
    },
    {
      "epoch": 0.05919707243569045,
      "grad_norm": 1.386881709098816,
      "learning_rate": 0.0005971102251912509,
      "loss": 5.5895,
      "step": 550
    },
    {
      "epoch": 0.06457862447529868,
      "grad_norm": 1.3578399419784546,
      "learning_rate": 0.0005967869841611895,
      "loss": 5.4912,
      "step": 600
    },
    {
      "epoch": 0.0699601765149069,
      "grad_norm": 1.2151751518249512,
      "learning_rate": 0.000596463743131128,
      "loss": 5.4275,
      "step": 650
    },
    {
      "epoch": 0.07534172855451512,
      "grad_norm": 1.3390365839004517,
      "learning_rate": 0.0005961405021010667,
      "loss": 5.3731,
      "step": 700
    },
    {
      "epoch": 0.08072328059412334,
      "grad_norm": 1.103965401649475,
      "learning_rate": 0.0005958172610710052,
      "loss": 5.3186,
      "step": 750
    },
    {
      "epoch": 0.08610483263373157,
      "grad_norm": 1.1866347789764404,
      "learning_rate": 0.0005954940200409439,
      "loss": 5.2687,
      "step": 800
    },
    {
      "epoch": 0.09148638467333979,
      "grad_norm": 1.0097291469573975,
      "learning_rate": 0.0005951707790108824,
      "loss": 5.2072,
      "step": 850
    },
    {
      "epoch": 0.09686793671294801,
      "grad_norm": 1.3843412399291992,
      "learning_rate": 0.0005948475379808209,
      "loss": 5.1549,
      "step": 900
    },
    {
      "epoch": 0.10224948875255624,
      "grad_norm": 1.1859126091003418,
      "learning_rate": 0.0005945242969507596,
      "loss": 5.0977,
      "step": 950
    },
    {
      "epoch": 0.10763104079216446,
      "grad_norm": 1.1825942993164062,
      "learning_rate": 0.0005942010559206981,
      "loss": 5.0644,
      "step": 1000
    },
    {
      "epoch": 0.10763104079216446,
      "eval_accuracy": 0.22798678911267642,
      "eval_loss": 5.015804290771484,
      "eval_runtime": 186.987,
      "eval_samples_per_second": 96.322,
      "eval_steps_per_second": 6.022,
      "step": 1000
    },
    {
      "epoch": 0.11301259283177269,
      "grad_norm": 1.036710262298584,
      "learning_rate": 0.0005938778148906367,
      "loss": 5.045,
      "step": 1050
    },
    {
      "epoch": 0.1183941448713809,
      "grad_norm": 1.1089870929718018,
      "learning_rate": 0.0005935545738605753,
      "loss": 5.0081,
      "step": 1100
    },
    {
      "epoch": 0.12377569691098914,
      "grad_norm": 0.9135867357254028,
      "learning_rate": 0.000593231332830514,
      "loss": 4.9925,
      "step": 1150
    },
    {
      "epoch": 0.12915724895059735,
      "grad_norm": 0.8327316641807556,
      "learning_rate": 0.0005929080918004525,
      "loss": 4.9437,
      "step": 1200
    },
    {
      "epoch": 0.13453880099020557,
      "grad_norm": 1.0137102603912354,
      "learning_rate": 0.000592584850770391,
      "loss": 4.9288,
      "step": 1250
    },
    {
      "epoch": 0.1399203530298138,
      "grad_norm": 1.620171070098877,
      "learning_rate": 0.0005922616097403296,
      "loss": 4.8813,
      "step": 1300
    },
    {
      "epoch": 0.14530190506942203,
      "grad_norm": 1.1357035636901855,
      "learning_rate": 0.0005919383687102682,
      "loss": 4.8713,
      "step": 1350
    },
    {
      "epoch": 0.15068345710903025,
      "grad_norm": 0.7975968718528748,
      "learning_rate": 0.0005916151276802069,
      "loss": 4.835,
      "step": 1400
    },
    {
      "epoch": 0.15606500914863847,
      "grad_norm": 0.9254979491233826,
      "learning_rate": 0.0005912918866501454,
      "loss": 4.806,
      "step": 1450
    },
    {
      "epoch": 0.16144656118824668,
      "grad_norm": 0.8283946514129639,
      "learning_rate": 0.000590968645620084,
      "loss": 4.8075,
      "step": 1500
    },
    {
      "epoch": 0.1668281132278549,
      "grad_norm": 1.1234453916549683,
      "learning_rate": 0.0005906454045900226,
      "loss": 4.7791,
      "step": 1550
    },
    {
      "epoch": 0.17220966526746315,
      "grad_norm": 0.9318310618400574,
      "learning_rate": 0.0005903221635599612,
      "loss": 4.7354,
      "step": 1600
    },
    {
      "epoch": 0.17759121730707136,
      "grad_norm": 0.8706691861152649,
      "learning_rate": 0.0005899989225298998,
      "loss": 4.7278,
      "step": 1650
    },
    {
      "epoch": 0.18297276934667958,
      "grad_norm": 1.0983690023422241,
      "learning_rate": 0.0005896756814998383,
      "loss": 4.7116,
      "step": 1700
    },
    {
      "epoch": 0.1883543213862878,
      "grad_norm": 1.1487178802490234,
      "learning_rate": 0.0005893524404697769,
      "loss": 4.7026,
      "step": 1750
    },
    {
      "epoch": 0.19373587342589602,
      "grad_norm": 0.8954223394393921,
      "learning_rate": 0.0005890291994397155,
      "loss": 4.6491,
      "step": 1800
    },
    {
      "epoch": 0.19911742546550426,
      "grad_norm": 1.1864057779312134,
      "learning_rate": 0.0005887059584096541,
      "loss": 4.6526,
      "step": 1850
    },
    {
      "epoch": 0.20449897750511248,
      "grad_norm": 0.945884108543396,
      "learning_rate": 0.0005883827173795926,
      "loss": 4.6132,
      "step": 1900
    },
    {
      "epoch": 0.2098805295447207,
      "grad_norm": 1.2601373195648193,
      "learning_rate": 0.0005880594763495313,
      "loss": 4.6044,
      "step": 1950
    },
    {
      "epoch": 0.2152620815843289,
      "grad_norm": 0.9447725415229797,
      "learning_rate": 0.0005877362353194698,
      "loss": 4.5877,
      "step": 2000
    },
    {
      "epoch": 0.2152620815843289,
      "eval_accuracy": 0.27040933128385014,
      "eval_loss": 4.508606910705566,
      "eval_runtime": 184.8398,
      "eval_samples_per_second": 97.441,
      "eval_steps_per_second": 6.092,
      "step": 2000
    },
    {
      "epoch": 0.22064363362393713,
      "grad_norm": 0.7851864695549011,
      "learning_rate": 0.0005874129942894084,
      "loss": 4.5643,
      "step": 2050
    },
    {
      "epoch": 0.22602518566354537,
      "grad_norm": 0.862130880355835,
      "learning_rate": 0.000587089753259347,
      "loss": 4.5481,
      "step": 2100
    },
    {
      "epoch": 0.2314067377031536,
      "grad_norm": 0.8470994830131531,
      "learning_rate": 0.0005867665122292855,
      "loss": 4.5219,
      "step": 2150
    },
    {
      "epoch": 0.2367882897427618,
      "grad_norm": 0.8713746070861816,
      "learning_rate": 0.0005864432711992242,
      "loss": 4.511,
      "step": 2200
    },
    {
      "epoch": 0.24216984178237003,
      "grad_norm": 0.8879591226577759,
      "learning_rate": 0.0005861200301691628,
      "loss": 4.5126,
      "step": 2250
    },
    {
      "epoch": 0.24755139382197827,
      "grad_norm": 0.9361649751663208,
      "learning_rate": 0.0005857967891391014,
      "loss": 4.4803,
      "step": 2300
    },
    {
      "epoch": 0.2529329458615865,
      "grad_norm": 0.8480615019798279,
      "learning_rate": 0.0005854735481090399,
      "loss": 4.4446,
      "step": 2350
    },
    {
      "epoch": 0.2583144979011947,
      "grad_norm": 1.0228937864303589,
      "learning_rate": 0.0005851503070789784,
      "loss": 4.4508,
      "step": 2400
    },
    {
      "epoch": 0.2636960499408029,
      "grad_norm": 0.9693055152893066,
      "learning_rate": 0.0005848270660489171,
      "loss": 4.4283,
      "step": 2450
    },
    {
      "epoch": 0.26907760198041114,
      "grad_norm": 0.7247486114501953,
      "learning_rate": 0.0005845038250188556,
      "loss": 4.4015,
      "step": 2500
    },
    {
      "epoch": 0.27445915402001936,
      "grad_norm": 0.7183290123939514,
      "learning_rate": 0.0005841805839887943,
      "loss": 4.3918,
      "step": 2550
    },
    {
      "epoch": 0.2798407060596276,
      "grad_norm": 0.900417149066925,
      "learning_rate": 0.0005838573429587328,
      "loss": 4.3701,
      "step": 2600
    },
    {
      "epoch": 0.2852222580992358,
      "grad_norm": 0.8292120099067688,
      "learning_rate": 0.0005835341019286715,
      "loss": 4.3891,
      "step": 2650
    },
    {
      "epoch": 0.29060381013884407,
      "grad_norm": 0.798037052154541,
      "learning_rate": 0.00058321086089861,
      "loss": 4.36,
      "step": 2700
    },
    {
      "epoch": 0.2959853621784523,
      "grad_norm": 0.8176091313362122,
      "learning_rate": 0.0005828876198685486,
      "loss": 4.3491,
      "step": 2750
    },
    {
      "epoch": 0.3013669142180605,
      "grad_norm": 0.8014965057373047,
      "learning_rate": 0.0005825643788384872,
      "loss": 4.3642,
      "step": 2800
    },
    {
      "epoch": 0.3067484662576687,
      "grad_norm": 1.2503150701522827,
      "learning_rate": 0.0005822411378084257,
      "loss": 4.3251,
      "step": 2850
    },
    {
      "epoch": 0.31213001829727693,
      "grad_norm": 0.7830126285552979,
      "learning_rate": 0.0005819178967783644,
      "loss": 4.3251,
      "step": 2900
    },
    {
      "epoch": 0.31751157033688515,
      "grad_norm": 0.9035398364067078,
      "learning_rate": 0.0005815946557483029,
      "loss": 4.3313,
      "step": 2950
    },
    {
      "epoch": 0.32289312237649337,
      "grad_norm": 0.6857808828353882,
      "learning_rate": 0.0005812714147182415,
      "loss": 4.2989,
      "step": 3000
    },
    {
      "epoch": 0.32289312237649337,
      "eval_accuracy": 0.2990024252411089,
      "eval_loss": 4.2293620109558105,
      "eval_runtime": 185.031,
      "eval_samples_per_second": 97.34,
      "eval_steps_per_second": 6.085,
      "step": 3000
    },
    {
      "epoch": 0.3282746744161016,
      "grad_norm": 0.7547754049301147,
      "learning_rate": 0.0005809481736881801,
      "loss": 4.296,
      "step": 3050
    },
    {
      "epoch": 0.3336562264557098,
      "grad_norm": 0.7557756304740906,
      "learning_rate": 0.0005806249326581187,
      "loss": 4.2866,
      "step": 3100
    },
    {
      "epoch": 0.3390377784953181,
      "grad_norm": 0.7948616147041321,
      "learning_rate": 0.0005803016916280573,
      "loss": 4.2736,
      "step": 3150
    },
    {
      "epoch": 0.3444193305349263,
      "grad_norm": 0.864937424659729,
      "learning_rate": 0.0005799784505979959,
      "loss": 4.2902,
      "step": 3200
    },
    {
      "epoch": 0.3498008825745345,
      "grad_norm": 0.8035250902175903,
      "learning_rate": 0.0005796552095679344,
      "loss": 4.2524,
      "step": 3250
    },
    {
      "epoch": 0.35518243461414273,
      "grad_norm": 0.7827973365783691,
      "learning_rate": 0.000579331968537873,
      "loss": 4.2573,
      "step": 3300
    },
    {
      "epoch": 0.36056398665375095,
      "grad_norm": 0.7145817875862122,
      "learning_rate": 0.0005790087275078116,
      "loss": 4.233,
      "step": 3350
    },
    {
      "epoch": 0.36594553869335916,
      "grad_norm": 0.7260717749595642,
      "learning_rate": 0.0005786854864777502,
      "loss": 4.2351,
      "step": 3400
    },
    {
      "epoch": 0.3713270907329674,
      "grad_norm": 0.846215009689331,
      "learning_rate": 0.0005783622454476888,
      "loss": 4.236,
      "step": 3450
    },
    {
      "epoch": 0.3767086427725756,
      "grad_norm": 0.6381611227989197,
      "learning_rate": 0.0005780390044176273,
      "loss": 4.2186,
      "step": 3500
    },
    {
      "epoch": 0.3820901948121838,
      "grad_norm": 0.7176727652549744,
      "learning_rate": 0.000577715763387566,
      "loss": 4.2159,
      "step": 3550
    },
    {
      "epoch": 0.38747174685179203,
      "grad_norm": 0.7199215888977051,
      "learning_rate": 0.0005773925223575045,
      "loss": 4.2187,
      "step": 3600
    },
    {
      "epoch": 0.3928532988914003,
      "grad_norm": 0.7120049595832825,
      "learning_rate": 0.0005770692813274432,
      "loss": 4.1858,
      "step": 3650
    },
    {
      "epoch": 0.3982348509310085,
      "grad_norm": 0.682905912399292,
      "learning_rate": 0.0005767460402973817,
      "loss": 4.2067,
      "step": 3700
    },
    {
      "epoch": 0.40361640297061674,
      "grad_norm": 0.8106050491333008,
      "learning_rate": 0.0005764227992673203,
      "loss": 4.1901,
      "step": 3750
    },
    {
      "epoch": 0.40899795501022496,
      "grad_norm": 0.6139677166938782,
      "learning_rate": 0.0005760995582372589,
      "loss": 4.1922,
      "step": 3800
    },
    {
      "epoch": 0.4143795070498332,
      "grad_norm": 0.6996195912361145,
      "learning_rate": 0.0005757763172071974,
      "loss": 4.1884,
      "step": 3850
    },
    {
      "epoch": 0.4197610590894414,
      "grad_norm": 0.6972913146018982,
      "learning_rate": 0.000575453076177136,
      "loss": 4.1469,
      "step": 3900
    },
    {
      "epoch": 0.4251426111290496,
      "grad_norm": 0.7805142998695374,
      "learning_rate": 0.0005751298351470746,
      "loss": 4.1681,
      "step": 3950
    },
    {
      "epoch": 0.4305241631686578,
      "grad_norm": 0.6750819683074951,
      "learning_rate": 0.0005748065941170133,
      "loss": 4.1677,
      "step": 4000
    },
    {
      "epoch": 0.4305241631686578,
      "eval_accuracy": 0.3131229545414788,
      "eval_loss": 4.085065841674805,
      "eval_runtime": 184.7939,
      "eval_samples_per_second": 97.465,
      "eval_steps_per_second": 6.093,
      "step": 4000
    },
    {
      "epoch": 0.43590571520826604,
      "grad_norm": 0.743445098400116,
      "learning_rate": 0.0005744833530869518,
      "loss": 4.1605,
      "step": 4050
    },
    {
      "epoch": 0.44128726724787426,
      "grad_norm": 0.6816270351409912,
      "learning_rate": 0.0005741601120568903,
      "loss": 4.1494,
      "step": 4100
    },
    {
      "epoch": 0.44666881928748253,
      "grad_norm": 0.6031078696250916,
      "learning_rate": 0.0005738368710268289,
      "loss": 4.1373,
      "step": 4150
    },
    {
      "epoch": 0.45205037132709075,
      "grad_norm": 0.869440495967865,
      "learning_rate": 0.0005735136299967675,
      "loss": 4.1274,
      "step": 4200
    },
    {
      "epoch": 0.45743192336669897,
      "grad_norm": 0.6965010762214661,
      "learning_rate": 0.0005731903889667062,
      "loss": 4.1363,
      "step": 4250
    },
    {
      "epoch": 0.4628134754063072,
      "grad_norm": 0.6586082577705383,
      "learning_rate": 0.0005728671479366447,
      "loss": 4.1073,
      "step": 4300
    },
    {
      "epoch": 0.4681950274459154,
      "grad_norm": 0.6008425951004028,
      "learning_rate": 0.0005725439069065833,
      "loss": 4.1131,
      "step": 4350
    },
    {
      "epoch": 0.4735765794855236,
      "grad_norm": 0.6233063340187073,
      "learning_rate": 0.0005722206658765219,
      "loss": 4.1266,
      "step": 4400
    },
    {
      "epoch": 0.47895813152513184,
      "grad_norm": 0.7806559205055237,
      "learning_rate": 0.0005718974248464605,
      "loss": 4.1052,
      "step": 4450
    },
    {
      "epoch": 0.48433968356474005,
      "grad_norm": 0.7604005336761475,
      "learning_rate": 0.000571574183816399,
      "loss": 4.0944,
      "step": 4500
    },
    {
      "epoch": 0.48972123560434827,
      "grad_norm": 0.8495274782180786,
      "learning_rate": 0.0005712509427863376,
      "loss": 4.1176,
      "step": 4550
    },
    {
      "epoch": 0.49510278764395654,
      "grad_norm": 0.6198326349258423,
      "learning_rate": 0.0005709277017562762,
      "loss": 4.1016,
      "step": 4600
    },
    {
      "epoch": 0.5004843396835648,
      "grad_norm": 0.6641437411308289,
      "learning_rate": 0.0005706044607262148,
      "loss": 4.0835,
      "step": 4650
    },
    {
      "epoch": 0.505865891723173,
      "grad_norm": 0.7691634297370911,
      "learning_rate": 0.0005702812196961534,
      "loss": 4.0815,
      "step": 4700
    },
    {
      "epoch": 0.5112474437627812,
      "grad_norm": 0.7156651020050049,
      "learning_rate": 0.0005699579786660919,
      "loss": 4.0818,
      "step": 4750
    },
    {
      "epoch": 0.5166289958023894,
      "grad_norm": 0.6472979784011841,
      "learning_rate": 0.0005696347376360306,
      "loss": 4.067,
      "step": 4800
    },
    {
      "epoch": 0.5220105478419976,
      "grad_norm": 0.724949300289154,
      "learning_rate": 0.0005693114966059691,
      "loss": 4.0615,
      "step": 4850
    },
    {
      "epoch": 0.5273920998816058,
      "grad_norm": 0.7559602856636047,
      "learning_rate": 0.0005689882555759077,
      "loss": 4.0676,
      "step": 4900
    },
    {
      "epoch": 0.5327736519212141,
      "grad_norm": 0.6394193768501282,
      "learning_rate": 0.0005686650145458463,
      "loss": 4.0461,
      "step": 4950
    },
    {
      "epoch": 0.5381552039608223,
      "grad_norm": 0.7001257538795471,
      "learning_rate": 0.0005683417735157848,
      "loss": 4.0729,
      "step": 5000
    },
    {
      "epoch": 0.5381552039608223,
      "eval_accuracy": 0.320991922635667,
      "eval_loss": 3.990946054458618,
      "eval_runtime": 184.8283,
      "eval_samples_per_second": 97.447,
      "eval_steps_per_second": 6.092,
      "step": 5000
    },
    {
      "epoch": 0.5435367560004305,
      "grad_norm": 0.7765357494354248,
      "learning_rate": 0.0005680185324857235,
      "loss": 4.0425,
      "step": 5050
    },
    {
      "epoch": 0.5489183080400387,
      "grad_norm": 0.6510383486747742,
      "learning_rate": 0.0005676952914556621,
      "loss": 4.0239,
      "step": 5100
    },
    {
      "epoch": 0.5542998600796469,
      "grad_norm": 0.6882658004760742,
      "learning_rate": 0.0005673720504256007,
      "loss": 4.0242,
      "step": 5150
    },
    {
      "epoch": 0.5596814121192552,
      "grad_norm": 0.7188264727592468,
      "learning_rate": 0.0005670488093955392,
      "loss": 4.0442,
      "step": 5200
    },
    {
      "epoch": 0.5650629641588634,
      "grad_norm": 0.6599631905555725,
      "learning_rate": 0.0005667255683654777,
      "loss": 4.0497,
      "step": 5250
    },
    {
      "epoch": 0.5704445161984716,
      "grad_norm": 0.5674473643302917,
      "learning_rate": 0.0005664023273354164,
      "loss": 4.0271,
      "step": 5300
    },
    {
      "epoch": 0.5758260682380799,
      "grad_norm": 0.6670402884483337,
      "learning_rate": 0.0005660790863053549,
      "loss": 4.0284,
      "step": 5350
    },
    {
      "epoch": 0.5812076202776881,
      "grad_norm": 0.7389664649963379,
      "learning_rate": 0.0005657558452752936,
      "loss": 4.0449,
      "step": 5400
    },
    {
      "epoch": 0.5865891723172963,
      "grad_norm": 0.6841723322868347,
      "learning_rate": 0.0005654326042452321,
      "loss": 4.051,
      "step": 5450
    },
    {
      "epoch": 0.5919707243569046,
      "grad_norm": 0.6119431853294373,
      "learning_rate": 0.0005651093632151708,
      "loss": 4.0158,
      "step": 5500
    },
    {
      "epoch": 0.5973522763965128,
      "grad_norm": 0.8850950598716736,
      "learning_rate": 0.0005647861221851093,
      "loss": 4.0183,
      "step": 5550
    },
    {
      "epoch": 0.602733828436121,
      "grad_norm": 0.6376111507415771,
      "learning_rate": 0.0005644628811550479,
      "loss": 4.0076,
      "step": 5600
    },
    {
      "epoch": 0.6081153804757292,
      "grad_norm": 0.5115376114845276,
      "learning_rate": 0.0005641396401249865,
      "loss": 4.0006,
      "step": 5650
    },
    {
      "epoch": 0.6134969325153374,
      "grad_norm": 0.6021561026573181,
      "learning_rate": 0.000563816399094925,
      "loss": 4.0056,
      "step": 5700
    },
    {
      "epoch": 0.6188784845549457,
      "grad_norm": 0.5795146226882935,
      "learning_rate": 0.0005634931580648637,
      "loss": 4.0159,
      "step": 5750
    },
    {
      "epoch": 0.6242600365945539,
      "grad_norm": 0.6918138861656189,
      "learning_rate": 0.0005631699170348022,
      "loss": 3.991,
      "step": 5800
    },
    {
      "epoch": 0.6296415886341621,
      "grad_norm": 0.6135892271995544,
      "learning_rate": 0.0005628466760047408,
      "loss": 4.0127,
      "step": 5850
    },
    {
      "epoch": 0.6350231406737703,
      "grad_norm": 0.6184191703796387,
      "learning_rate": 0.0005625234349746794,
      "loss": 3.9862,
      "step": 5900
    },
    {
      "epoch": 0.6404046927133785,
      "grad_norm": 0.6259024739265442,
      "learning_rate": 0.000562200193944618,
      "loss": 3.9922,
      "step": 5950
    },
    {
      "epoch": 0.6457862447529867,
      "grad_norm": 0.5602654814720154,
      "learning_rate": 0.0005618769529145566,
      "loss": 3.995,
      "step": 6000
    },
    {
      "epoch": 0.6457862447529867,
      "eval_accuracy": 0.32913100180896193,
      "eval_loss": 3.9143569469451904,
      "eval_runtime": 184.5106,
      "eval_samples_per_second": 97.615,
      "eval_steps_per_second": 6.103,
      "step": 6000
    },
    {
      "epoch": 0.651167796792595,
      "grad_norm": 0.7381386160850525,
      "learning_rate": 0.0005615537118844952,
      "loss": 3.9749,
      "step": 6050
    },
    {
      "epoch": 0.6565493488322032,
      "grad_norm": 0.6130694150924683,
      "learning_rate": 0.0005612304708544337,
      "loss": 3.9791,
      "step": 6100
    },
    {
      "epoch": 0.6619309008718114,
      "grad_norm": 0.6282258629798889,
      "learning_rate": 0.0005609072298243723,
      "loss": 3.9766,
      "step": 6150
    },
    {
      "epoch": 0.6673124529114196,
      "grad_norm": 0.7320833802223206,
      "learning_rate": 0.0005605839887943109,
      "loss": 3.9744,
      "step": 6200
    },
    {
      "epoch": 0.6726940049510278,
      "grad_norm": 0.7067676782608032,
      "learning_rate": 0.0005602607477642495,
      "loss": 3.9755,
      "step": 6250
    },
    {
      "epoch": 0.6780755569906362,
      "grad_norm": 0.6241488456726074,
      "learning_rate": 0.0005599375067341881,
      "loss": 3.9639,
      "step": 6300
    },
    {
      "epoch": 0.6834571090302444,
      "grad_norm": 0.6031643152236938,
      "learning_rate": 0.0005596142657041266,
      "loss": 3.9535,
      "step": 6350
    },
    {
      "epoch": 0.6888386610698526,
      "grad_norm": 0.5485600829124451,
      "learning_rate": 0.0005592910246740653,
      "loss": 3.9616,
      "step": 6400
    },
    {
      "epoch": 0.6942202131094608,
      "grad_norm": 0.6337101459503174,
      "learning_rate": 0.0005589677836440038,
      "loss": 3.9509,
      "step": 6450
    },
    {
      "epoch": 0.699601765149069,
      "grad_norm": 0.546737790107727,
      "learning_rate": 0.0005586445426139425,
      "loss": 3.9416,
      "step": 6500
    },
    {
      "epoch": 0.7049833171886772,
      "grad_norm": 0.6616541147232056,
      "learning_rate": 0.000558321301583881,
      "loss": 3.9618,
      "step": 6550
    },
    {
      "epoch": 0.7103648692282855,
      "grad_norm": 0.5739938020706177,
      "learning_rate": 0.0005579980605538196,
      "loss": 3.949,
      "step": 6600
    },
    {
      "epoch": 0.7157464212678937,
      "grad_norm": 0.5481446981430054,
      "learning_rate": 0.0005576748195237582,
      "loss": 3.961,
      "step": 6650
    },
    {
      "epoch": 0.7211279733075019,
      "grad_norm": 0.6824391484260559,
      "learning_rate": 0.0005573515784936967,
      "loss": 3.9376,
      "step": 6700
    },
    {
      "epoch": 0.7265095253471101,
      "grad_norm": 0.6663408279418945,
      "learning_rate": 0.0005570283374636353,
      "loss": 3.9525,
      "step": 6750
    },
    {
      "epoch": 0.7318910773867183,
      "grad_norm": 0.5435226559638977,
      "learning_rate": 0.0005567050964335739,
      "loss": 3.9249,
      "step": 6800
    },
    {
      "epoch": 0.7372726294263265,
      "grad_norm": 0.6284464001655579,
      "learning_rate": 0.0005563818554035126,
      "loss": 3.9307,
      "step": 6850
    },
    {
      "epoch": 0.7426541814659348,
      "grad_norm": 0.5826367139816284,
      "learning_rate": 0.0005560586143734511,
      "loss": 3.9251,
      "step": 6900
    },
    {
      "epoch": 0.748035733505543,
      "grad_norm": 0.5755831003189087,
      "learning_rate": 0.0005557353733433896,
      "loss": 3.9271,
      "step": 6950
    },
    {
      "epoch": 0.7534172855451512,
      "grad_norm": 0.6272289156913757,
      "learning_rate": 0.0005554121323133283,
      "loss": 3.9517,
      "step": 7000
    },
    {
      "epoch": 0.7534172855451512,
      "eval_accuracy": 0.33305206722441094,
      "eval_loss": 3.8637194633483887,
      "eval_runtime": 184.9315,
      "eval_samples_per_second": 97.393,
      "eval_steps_per_second": 6.089,
      "step": 7000
    },
    {
      "epoch": 0.7587988375847594,
      "grad_norm": 0.6227930188179016,
      "learning_rate": 0.0005550888912832668,
      "loss": 3.916,
      "step": 7050
    },
    {
      "epoch": 0.7641803896243676,
      "grad_norm": 0.6726216077804565,
      "learning_rate": 0.0005547656502532055,
      "loss": 3.9217,
      "step": 7100
    },
    {
      "epoch": 0.7695619416639758,
      "grad_norm": 0.5420123338699341,
      "learning_rate": 0.000554442409223144,
      "loss": 3.9238,
      "step": 7150
    },
    {
      "epoch": 0.7749434937035841,
      "grad_norm": 0.5879136919975281,
      "learning_rate": 0.0005541191681930826,
      "loss": 3.9191,
      "step": 7200
    },
    {
      "epoch": 0.7803250457431924,
      "grad_norm": 0.5138105750083923,
      "learning_rate": 0.0005537959271630212,
      "loss": 3.9145,
      "step": 7250
    },
    {
      "epoch": 0.7857065977828006,
      "grad_norm": 0.5751107931137085,
      "learning_rate": 0.0005534726861329598,
      "loss": 3.9067,
      "step": 7300
    },
    {
      "epoch": 0.7910881498224088,
      "grad_norm": 0.6342102289199829,
      "learning_rate": 0.0005531494451028983,
      "loss": 3.9038,
      "step": 7350
    },
    {
      "epoch": 0.796469701862017,
      "grad_norm": 0.5363113284111023,
      "learning_rate": 0.0005528262040728369,
      "loss": 3.9121,
      "step": 7400
    },
    {
      "epoch": 0.8018512539016253,
      "grad_norm": 0.6412177681922913,
      "learning_rate": 0.0005525029630427755,
      "loss": 3.894,
      "step": 7450
    },
    {
      "epoch": 0.8072328059412335,
      "grad_norm": 0.6429057121276855,
      "learning_rate": 0.0005521797220127141,
      "loss": 3.904,
      "step": 7500
    },
    {
      "epoch": 0.8126143579808417,
      "grad_norm": 0.6362873315811157,
      "learning_rate": 0.0005518564809826527,
      "loss": 3.9328,
      "step": 7550
    },
    {
      "epoch": 0.8179959100204499,
      "grad_norm": 0.6349565386772156,
      "learning_rate": 0.0005515332399525912,
      "loss": 3.8807,
      "step": 7600
    },
    {
      "epoch": 0.8233774620600581,
      "grad_norm": 0.6769669651985168,
      "learning_rate": 0.0005512099989225299,
      "loss": 3.8932,
      "step": 7650
    },
    {
      "epoch": 0.8287590140996663,
      "grad_norm": 0.601761519908905,
      "learning_rate": 0.0005508867578924685,
      "loss": 3.897,
      "step": 7700
    },
    {
      "epoch": 0.8341405661392746,
      "grad_norm": 0.6315547823905945,
      "learning_rate": 0.000550563516862407,
      "loss": 3.8814,
      "step": 7750
    },
    {
      "epoch": 0.8395221181788828,
      "grad_norm": 0.5754820108413696,
      "learning_rate": 0.0005502402758323456,
      "loss": 3.9101,
      "step": 7800
    },
    {
      "epoch": 0.844903670218491,
      "grad_norm": 0.627307653427124,
      "learning_rate": 0.0005499170348022841,
      "loss": 3.9035,
      "step": 7850
    },
    {
      "epoch": 0.8502852222580992,
      "grad_norm": 0.6498885154724121,
      "learning_rate": 0.0005495937937722228,
      "loss": 3.9012,
      "step": 7900
    },
    {
      "epoch": 0.8556667742977074,
      "grad_norm": 0.5503144264221191,
      "learning_rate": 0.0005492705527421614,
      "loss": 3.8942,
      "step": 7950
    },
    {
      "epoch": 0.8610483263373157,
      "grad_norm": 0.6200529336929321,
      "learning_rate": 0.0005489473117121,
      "loss": 3.8717,
      "step": 8000
    },
    {
      "epoch": 0.8610483263373157,
      "eval_accuracy": 0.3380203291726159,
      "eval_loss": 3.8143210411071777,
      "eval_runtime": 184.8318,
      "eval_samples_per_second": 97.445,
      "eval_steps_per_second": 6.092,
      "step": 8000
    },
    {
      "epoch": 0.8664298783769239,
      "grad_norm": 0.7157406210899353,
      "learning_rate": 0.0005486240706820385,
      "loss": 3.8746,
      "step": 8050
    },
    {
      "epoch": 0.8718114304165321,
      "grad_norm": 0.643976628780365,
      "learning_rate": 0.0005483008296519772,
      "loss": 3.8859,
      "step": 8100
    },
    {
      "epoch": 0.8771929824561403,
      "grad_norm": 0.5374308824539185,
      "learning_rate": 0.0005479775886219157,
      "loss": 3.8586,
      "step": 8150
    },
    {
      "epoch": 0.8825745344957485,
      "grad_norm": 0.6169512867927551,
      "learning_rate": 0.0005476543475918542,
      "loss": 3.8734,
      "step": 8200
    },
    {
      "epoch": 0.8879560865353568,
      "grad_norm": 0.6148831248283386,
      "learning_rate": 0.0005473311065617929,
      "loss": 3.8731,
      "step": 8250
    },
    {
      "epoch": 0.8933376385749651,
      "grad_norm": 0.5680239796638489,
      "learning_rate": 0.0005470078655317314,
      "loss": 3.8674,
      "step": 8300
    },
    {
      "epoch": 0.8987191906145733,
      "grad_norm": 0.5679119825363159,
      "learning_rate": 0.0005466846245016701,
      "loss": 3.8556,
      "step": 8350
    },
    {
      "epoch": 0.9041007426541815,
      "grad_norm": 0.5395457744598389,
      "learning_rate": 0.0005463613834716086,
      "loss": 3.8536,
      "step": 8400
    },
    {
      "epoch": 0.9094822946937897,
      "grad_norm": 0.5603843331336975,
      "learning_rate": 0.0005460381424415472,
      "loss": 3.8579,
      "step": 8450
    },
    {
      "epoch": 0.9148638467333979,
      "grad_norm": 0.5550517439842224,
      "learning_rate": 0.0005457149014114858,
      "loss": 3.858,
      "step": 8500
    },
    {
      "epoch": 0.9202453987730062,
      "grad_norm": 0.6112827062606812,
      "learning_rate": 0.0005453916603814243,
      "loss": 3.8563,
      "step": 8550
    },
    {
      "epoch": 0.9256269508126144,
      "grad_norm": 0.6317242383956909,
      "learning_rate": 0.000545068419351363,
      "loss": 3.854,
      "step": 8600
    },
    {
      "epoch": 0.9310085028522226,
      "grad_norm": 0.574309766292572,
      "learning_rate": 0.0005447451783213015,
      "loss": 3.8455,
      "step": 8650
    },
    {
      "epoch": 0.9363900548918308,
      "grad_norm": 0.5467745065689087,
      "learning_rate": 0.0005444219372912401,
      "loss": 3.858,
      "step": 8700
    },
    {
      "epoch": 0.941771606931439,
      "grad_norm": 0.5530813932418823,
      "learning_rate": 0.0005440986962611787,
      "loss": 3.8518,
      "step": 8750
    },
    {
      "epoch": 0.9471531589710472,
      "grad_norm": 0.7071423530578613,
      "learning_rate": 0.0005437754552311173,
      "loss": 3.8473,
      "step": 8800
    },
    {
      "epoch": 0.9525347110106555,
      "grad_norm": 0.5127449631690979,
      "learning_rate": 0.0005434522142010559,
      "loss": 3.8471,
      "step": 8850
    },
    {
      "epoch": 0.9579162630502637,
      "grad_norm": 0.5855455994606018,
      "learning_rate": 0.0005431289731709945,
      "loss": 3.8589,
      "step": 8900
    },
    {
      "epoch": 0.9632978150898719,
      "grad_norm": 0.6145156025886536,
      "learning_rate": 0.000542805732140933,
      "loss": 3.8295,
      "step": 8950
    },
    {
      "epoch": 0.9686793671294801,
      "grad_norm": 0.5550306439399719,
      "learning_rate": 0.0005424824911108716,
      "loss": 3.8378,
      "step": 9000
    },
    {
      "epoch": 0.9686793671294801,
      "eval_accuracy": 0.3415591537287335,
      "eval_loss": 3.7806262969970703,
      "eval_runtime": 184.435,
      "eval_samples_per_second": 97.655,
      "eval_steps_per_second": 6.105,
      "step": 9000
    },
    {
      "epoch": 0.9740609191690883,
      "grad_norm": 0.5801270008087158,
      "learning_rate": 0.0005421592500808102,
      "loss": 3.8502,
      "step": 9050
    },
    {
      "epoch": 0.9794424712086965,
      "grad_norm": 0.5607963800430298,
      "learning_rate": 0.0005418360090507488,
      "loss": 3.825,
      "step": 9100
    },
    {
      "epoch": 0.9848240232483048,
      "grad_norm": 0.5534636974334717,
      "learning_rate": 0.0005415127680206874,
      "loss": 3.8361,
      "step": 9150
    },
    {
      "epoch": 0.9902055752879131,
      "grad_norm": 0.692423939704895,
      "learning_rate": 0.0005411895269906259,
      "loss": 3.8318,
      "step": 9200
    },
    {
      "epoch": 0.9955871273275213,
      "grad_norm": 0.5722038745880127,
      "learning_rate": 0.0005408662859605646,
      "loss": 3.8376,
      "step": 9250
    },
    {
      "epoch": 1.0009686793671295,
      "grad_norm": 0.6226180195808411,
      "learning_rate": 0.0005405430449305031,
      "loss": 3.8294,
      "step": 9300
    },
    {
      "epoch": 1.0063502314067376,
      "grad_norm": 0.6096346974372864,
      "learning_rate": 0.0005402198039004416,
      "loss": 3.7878,
      "step": 9350
    },
    {
      "epoch": 1.011731783446346,
      "grad_norm": 0.5371617078781128,
      "learning_rate": 0.0005398965628703803,
      "loss": 3.7627,
      "step": 9400
    },
    {
      "epoch": 1.017113335485954,
      "grad_norm": 0.6130861639976501,
      "learning_rate": 0.0005395733218403189,
      "loss": 3.769,
      "step": 9450
    },
    {
      "epoch": 1.0224948875255624,
      "grad_norm": 0.5652937889099121,
      "learning_rate": 0.0005392500808102575,
      "loss": 3.7714,
      "step": 9500
    },
    {
      "epoch": 1.0278764395651705,
      "grad_norm": 0.5810158252716064,
      "learning_rate": 0.000538926839780196,
      "loss": 3.7566,
      "step": 9550
    },
    {
      "epoch": 1.0332579916047788,
      "grad_norm": 0.6366006731987,
      "learning_rate": 0.0005386035987501346,
      "loss": 3.7647,
      "step": 9600
    },
    {
      "epoch": 1.0386395436443872,
      "grad_norm": 0.5298298597335815,
      "learning_rate": 0.0005382803577200732,
      "loss": 3.7488,
      "step": 9650
    },
    {
      "epoch": 1.0440210956839953,
      "grad_norm": 1.2535181045532227,
      "learning_rate": 0.0005379571166900119,
      "loss": 3.7539,
      "step": 9700
    },
    {
      "epoch": 1.0494026477236036,
      "grad_norm": 0.601189374923706,
      "learning_rate": 0.0005376338756599504,
      "loss": 3.7654,
      "step": 9750
    },
    {
      "epoch": 1.0547841997632117,
      "grad_norm": 0.6602137684822083,
      "learning_rate": 0.0005373106346298889,
      "loss": 3.7686,
      "step": 9800
    },
    {
      "epoch": 1.06016575180282,
      "grad_norm": 0.6176168918609619,
      "learning_rate": 0.0005369873935998276,
      "loss": 3.7589,
      "step": 9850
    },
    {
      "epoch": 1.0655473038424281,
      "grad_norm": 0.6260287165641785,
      "learning_rate": 0.0005366641525697661,
      "loss": 3.7655,
      "step": 9900
    },
    {
      "epoch": 1.0709288558820365,
      "grad_norm": 0.5720372796058655,
      "learning_rate": 0.0005363409115397048,
      "loss": 3.7559,
      "step": 9950
    },
    {
      "epoch": 1.0763104079216446,
      "grad_norm": 0.5716636180877686,
      "learning_rate": 0.0005360176705096433,
      "loss": 3.7573,
      "step": 10000
    },
    {
      "epoch": 1.0763104079216446,
      "eval_accuracy": 0.3451164492757796,
      "eval_loss": 3.7458271980285645,
      "eval_runtime": 185.1344,
      "eval_samples_per_second": 97.286,
      "eval_steps_per_second": 6.082,
      "step": 10000
    },
    {
      "epoch": 1.081691959961253,
      "grad_norm": 0.6392249464988708,
      "learning_rate": 0.0005356944294795819,
      "loss": 3.746,
      "step": 10050
    },
    {
      "epoch": 1.087073512000861,
      "grad_norm": 0.5700231194496155,
      "learning_rate": 0.0005353711884495205,
      "loss": 3.7913,
      "step": 10100
    },
    {
      "epoch": 1.0924550640404693,
      "grad_norm": 0.5578026175498962,
      "learning_rate": 0.000535047947419459,
      "loss": 3.7517,
      "step": 10150
    },
    {
      "epoch": 1.0978366160800774,
      "grad_norm": 0.5823580026626587,
      "learning_rate": 0.0005347247063893976,
      "loss": 3.7495,
      "step": 10200
    },
    {
      "epoch": 1.1032181681196858,
      "grad_norm": 0.6097853779792786,
      "learning_rate": 0.0005344014653593362,
      "loss": 3.748,
      "step": 10250
    },
    {
      "epoch": 1.1085997201592939,
      "grad_norm": 0.6023664474487305,
      "learning_rate": 0.0005340782243292748,
      "loss": 3.7464,
      "step": 10300
    },
    {
      "epoch": 1.1139812721989022,
      "grad_norm": 0.6338013410568237,
      "learning_rate": 0.0005337549832992134,
      "loss": 3.7546,
      "step": 10350
    },
    {
      "epoch": 1.1193628242385103,
      "grad_norm": 0.5424166917800903,
      "learning_rate": 0.000533431742269152,
      "loss": 3.7676,
      "step": 10400
    },
    {
      "epoch": 1.1247443762781186,
      "grad_norm": 0.5553328990936279,
      "learning_rate": 0.0005331085012390905,
      "loss": 3.7516,
      "step": 10450
    },
    {
      "epoch": 1.1301259283177267,
      "grad_norm": 0.5472305417060852,
      "learning_rate": 0.0005327852602090292,
      "loss": 3.7578,
      "step": 10500
    },
    {
      "epoch": 1.135507480357335,
      "grad_norm": 0.6304797530174255,
      "learning_rate": 0.0005324620191789678,
      "loss": 3.7302,
      "step": 10550
    },
    {
      "epoch": 1.1408890323969434,
      "grad_norm": 0.5144755244255066,
      "learning_rate": 0.0005321387781489063,
      "loss": 3.7487,
      "step": 10600
    },
    {
      "epoch": 1.1462705844365515,
      "grad_norm": 0.5785931348800659,
      "learning_rate": 0.0005318155371188449,
      "loss": 3.7557,
      "step": 10650
    },
    {
      "epoch": 1.1516521364761596,
      "grad_norm": 0.5502986907958984,
      "learning_rate": 0.0005314922960887834,
      "loss": 3.7452,
      "step": 10700
    },
    {
      "epoch": 1.157033688515768,
      "grad_norm": 0.6079390048980713,
      "learning_rate": 0.0005311690550587221,
      "loss": 3.7587,
      "step": 10750
    },
    {
      "epoch": 1.1624152405553763,
      "grad_norm": 0.5040614008903503,
      "learning_rate": 0.0005308458140286607,
      "loss": 3.7611,
      "step": 10800
    },
    {
      "epoch": 1.1677967925949844,
      "grad_norm": 0.5929045677185059,
      "learning_rate": 0.0005305225729985993,
      "loss": 3.7669,
      "step": 10850
    },
    {
      "epoch": 1.1731783446345927,
      "grad_norm": 0.535692572593689,
      "learning_rate": 0.0005301993319685378,
      "loss": 3.764,
      "step": 10900
    },
    {
      "epoch": 1.1785598966742008,
      "grad_norm": 0.5377613306045532,
      "learning_rate": 0.0005298760909384765,
      "loss": 3.7536,
      "step": 10950
    },
    {
      "epoch": 1.1839414487138091,
      "grad_norm": 0.6084091663360596,
      "learning_rate": 0.000529552849908415,
      "loss": 3.7359,
      "step": 11000
    },
    {
      "epoch": 1.1839414487138091,
      "eval_accuracy": 0.3476050350182825,
      "eval_loss": 3.722259044647217,
      "eval_runtime": 184.9277,
      "eval_samples_per_second": 97.395,
      "eval_steps_per_second": 6.089,
      "step": 11000
    },
    {
      "epoch": 1.1893230007534172,
      "grad_norm": 0.5570613741874695,
      "learning_rate": 0.0005292296088783535,
      "loss": 3.7553,
      "step": 11050
    },
    {
      "epoch": 1.1947045527930256,
      "grad_norm": 0.5923260450363159,
      "learning_rate": 0.0005289063678482922,
      "loss": 3.7489,
      "step": 11100
    },
    {
      "epoch": 1.2000861048326337,
      "grad_norm": 0.6002750396728516,
      "learning_rate": 0.0005285831268182307,
      "loss": 3.742,
      "step": 11150
    },
    {
      "epoch": 1.205467656872242,
      "grad_norm": 0.6400654911994934,
      "learning_rate": 0.0005282598857881694,
      "loss": 3.7462,
      "step": 11200
    },
    {
      "epoch": 1.21084920891185,
      "grad_norm": 0.5764461755752563,
      "learning_rate": 0.0005279366447581079,
      "loss": 3.7248,
      "step": 11250
    },
    {
      "epoch": 1.2162307609514584,
      "grad_norm": 0.5291248559951782,
      "learning_rate": 0.0005276134037280465,
      "loss": 3.7395,
      "step": 11300
    },
    {
      "epoch": 1.2216123129910665,
      "grad_norm": 0.5363642573356628,
      "learning_rate": 0.0005272966275185863,
      "loss": 3.7486,
      "step": 11350
    },
    {
      "epoch": 1.2269938650306749,
      "grad_norm": 0.5614747405052185,
      "learning_rate": 0.0005269733864885249,
      "loss": 3.7352,
      "step": 11400
    },
    {
      "epoch": 1.232375417070283,
      "grad_norm": 0.568588137626648,
      "learning_rate": 0.0005266501454584636,
      "loss": 3.7354,
      "step": 11450
    },
    {
      "epoch": 1.2377569691098913,
      "grad_norm": 0.5658771991729736,
      "learning_rate": 0.0005263269044284021,
      "loss": 3.7457,
      "step": 11500
    },
    {
      "epoch": 1.2431385211494996,
      "grad_norm": 0.5699712038040161,
      "learning_rate": 0.0005260036633983406,
      "loss": 3.7583,
      "step": 11550
    },
    {
      "epoch": 1.2485200731891077,
      "grad_norm": 0.6179584264755249,
      "learning_rate": 0.0005256804223682792,
      "loss": 3.7221,
      "step": 11600
    },
    {
      "epoch": 1.2539016252287158,
      "grad_norm": 0.5554487109184265,
      "learning_rate": 0.0005253571813382178,
      "loss": 3.7318,
      "step": 11650
    },
    {
      "epoch": 1.2592831772683242,
      "grad_norm": 0.5801360011100769,
      "learning_rate": 0.0005250339403081564,
      "loss": 3.7272,
      "step": 11700
    },
    {
      "epoch": 1.2646647293079325,
      "grad_norm": 0.58757084608078,
      "learning_rate": 0.000524710699278095,
      "loss": 3.7316,
      "step": 11750
    },
    {
      "epoch": 1.2700462813475406,
      "grad_norm": 0.6320570111274719,
      "learning_rate": 0.0005243874582480336,
      "loss": 3.7323,
      "step": 11800
    },
    {
      "epoch": 1.275427833387149,
      "grad_norm": 0.5550466179847717,
      "learning_rate": 0.0005240642172179722,
      "loss": 3.7152,
      "step": 11850
    },
    {
      "epoch": 1.280809385426757,
      "grad_norm": 0.7199887633323669,
      "learning_rate": 0.0005237409761879107,
      "loss": 3.724,
      "step": 11900
    },
    {
      "epoch": 1.2861909374663654,
      "grad_norm": 0.5186243653297424,
      "learning_rate": 0.0005234177351578493,
      "loss": 3.7295,
      "step": 11950
    },
    {
      "epoch": 1.2915724895059735,
      "grad_norm": 0.5249941945075989,
      "learning_rate": 0.0005230944941277878,
      "loss": 3.7189,
      "step": 12000
    },
    {
      "epoch": 1.2915724895059735,
      "eval_accuracy": 0.34949287894405906,
      "eval_loss": 3.6969494819641113,
      "eval_runtime": 184.7681,
      "eval_samples_per_second": 97.479,
      "eval_steps_per_second": 6.094,
      "step": 12000
    },
    {
      "epoch": 1.2969540415455818,
      "grad_norm": 0.5758280158042908,
      "learning_rate": 0.0005227712530977265,
      "loss": 3.7271,
      "step": 12050
    },
    {
      "epoch": 1.30233559358519,
      "grad_norm": 0.5738679766654968,
      "learning_rate": 0.0005224480120676651,
      "loss": 3.7388,
      "step": 12100
    },
    {
      "epoch": 1.3077171456247982,
      "grad_norm": 0.6011477708816528,
      "learning_rate": 0.0005221247710376037,
      "loss": 3.7381,
      "step": 12150
    },
    {
      "epoch": 1.3130986976644063,
      "grad_norm": 0.5592882037162781,
      "learning_rate": 0.0005218015300075422,
      "loss": 3.7326,
      "step": 12200
    },
    {
      "epoch": 1.3184802497040147,
      "grad_norm": 0.5867257714271545,
      "learning_rate": 0.0005214782889774809,
      "loss": 3.7186,
      "step": 12250
    },
    {
      "epoch": 1.3238618017436228,
      "grad_norm": 0.5920559167861938,
      "learning_rate": 0.0005211550479474194,
      "loss": 3.7018,
      "step": 12300
    },
    {
      "epoch": 1.329243353783231,
      "grad_norm": 0.5726664066314697,
      "learning_rate": 0.000520831806917358,
      "loss": 3.7209,
      "step": 12350
    },
    {
      "epoch": 1.3346249058228392,
      "grad_norm": 0.5191407799720764,
      "learning_rate": 0.0005205085658872966,
      "loss": 3.7221,
      "step": 12400
    },
    {
      "epoch": 1.3400064578624475,
      "grad_norm": 0.5495452284812927,
      "learning_rate": 0.0005201853248572351,
      "loss": 3.7186,
      "step": 12450
    },
    {
      "epoch": 1.3453880099020559,
      "grad_norm": 0.536133348941803,
      "learning_rate": 0.0005198620838271738,
      "loss": 3.727,
      "step": 12500
    },
    {
      "epoch": 1.350769561941664,
      "grad_norm": 0.6578084826469421,
      "learning_rate": 0.0005195388427971123,
      "loss": 3.7175,
      "step": 12550
    },
    {
      "epoch": 1.356151113981272,
      "grad_norm": 0.697626531124115,
      "learning_rate": 0.000519215601767051,
      "loss": 3.7129,
      "step": 12600
    },
    {
      "epoch": 1.3615326660208804,
      "grad_norm": 0.5493432879447937,
      "learning_rate": 0.0005188923607369895,
      "loss": 3.7113,
      "step": 12650
    },
    {
      "epoch": 1.3669142180604887,
      "grad_norm": 0.5725867748260498,
      "learning_rate": 0.000518569119706928,
      "loss": 3.7209,
      "step": 12700
    },
    {
      "epoch": 1.3722957701000968,
      "grad_norm": 0.5644043684005737,
      "learning_rate": 0.0005182458786768667,
      "loss": 3.7148,
      "step": 12750
    },
    {
      "epoch": 1.3776773221397052,
      "grad_norm": 0.516650378704071,
      "learning_rate": 0.0005179226376468052,
      "loss": 3.7039,
      "step": 12800
    },
    {
      "epoch": 1.3830588741793133,
      "grad_norm": 0.5859173536300659,
      "learning_rate": 0.0005175993966167438,
      "loss": 3.7156,
      "step": 12850
    },
    {
      "epoch": 1.3884404262189216,
      "grad_norm": 0.4950462579727173,
      "learning_rate": 0.0005172761555866824,
      "loss": 3.7052,
      "step": 12900
    },
    {
      "epoch": 1.3938219782585297,
      "grad_norm": 0.6020005941390991,
      "learning_rate": 0.0005169529145566211,
      "loss": 3.7147,
      "step": 12950
    },
    {
      "epoch": 1.399203530298138,
      "grad_norm": 0.5504841208457947,
      "learning_rate": 0.0005166296735265596,
      "loss": 3.7161,
      "step": 13000
    },
    {
      "epoch": 1.399203530298138,
      "eval_accuracy": 0.3519344179861383,
      "eval_loss": 3.6728596687316895,
      "eval_runtime": 184.9723,
      "eval_samples_per_second": 97.371,
      "eval_steps_per_second": 6.087,
      "step": 13000
    },
    {
      "epoch": 1.4045850823377461,
      "grad_norm": 0.5823192000389099,
      "learning_rate": 0.0005163064324964982,
      "loss": 3.7071,
      "step": 13050
    },
    {
      "epoch": 1.4099666343773545,
      "grad_norm": 0.5856388807296753,
      "learning_rate": 0.0005159831914664367,
      "loss": 3.6995,
      "step": 13100
    },
    {
      "epoch": 1.4153481864169626,
      "grad_norm": 0.6325433850288391,
      "learning_rate": 0.0005156599504363753,
      "loss": 3.717,
      "step": 13150
    },
    {
      "epoch": 1.420729738456571,
      "grad_norm": 0.6164390444755554,
      "learning_rate": 0.000515336709406314,
      "loss": 3.7094,
      "step": 13200
    },
    {
      "epoch": 1.426111290496179,
      "grad_norm": 0.5653883218765259,
      "learning_rate": 0.0005150134683762525,
      "loss": 3.6838,
      "step": 13250
    },
    {
      "epoch": 1.4314928425357873,
      "grad_norm": 0.6306009888648987,
      "learning_rate": 0.0005146902273461911,
      "loss": 3.7257,
      "step": 13300
    },
    {
      "epoch": 1.4368743945753955,
      "grad_norm": 0.5829269886016846,
      "learning_rate": 0.0005143669863161297,
      "loss": 3.7158,
      "step": 13350
    },
    {
      "epoch": 1.4422559466150038,
      "grad_norm": 0.5895189642906189,
      "learning_rate": 0.0005140437452860683,
      "loss": 3.7031,
      "step": 13400
    },
    {
      "epoch": 1.447637498654612,
      "grad_norm": 0.5675577521324158,
      "learning_rate": 0.0005137205042560069,
      "loss": 3.7066,
      "step": 13450
    },
    {
      "epoch": 1.4530190506942202,
      "grad_norm": 0.5363256335258484,
      "learning_rate": 0.0005133972632259455,
      "loss": 3.6979,
      "step": 13500
    },
    {
      "epoch": 1.4584006027338283,
      "grad_norm": 0.6230180859565735,
      "learning_rate": 0.000513074022195884,
      "loss": 3.7036,
      "step": 13550
    },
    {
      "epoch": 1.4637821547734367,
      "grad_norm": 0.566116988658905,
      "learning_rate": 0.0005127507811658226,
      "loss": 3.7168,
      "step": 13600
    },
    {
      "epoch": 1.469163706813045,
      "grad_norm": 0.5180698037147522,
      "learning_rate": 0.0005124275401357612,
      "loss": 3.7075,
      "step": 13650
    },
    {
      "epoch": 1.474545258852653,
      "grad_norm": 0.743535578250885,
      "learning_rate": 0.0005121042991056997,
      "loss": 3.7021,
      "step": 13700
    },
    {
      "epoch": 1.4799268108922612,
      "grad_norm": 0.5717049241065979,
      "learning_rate": 0.0005117810580756384,
      "loss": 3.7071,
      "step": 13750
    },
    {
      "epoch": 1.4853083629318695,
      "grad_norm": 0.5536872148513794,
      "learning_rate": 0.0005114578170455769,
      "loss": 3.7012,
      "step": 13800
    },
    {
      "epoch": 1.4906899149714778,
      "grad_norm": 0.5160924196243286,
      "learning_rate": 0.0005111345760155156,
      "loss": 3.7051,
      "step": 13850
    },
    {
      "epoch": 1.496071467011086,
      "grad_norm": 0.5639503598213196,
      "learning_rate": 0.0005108113349854541,
      "loss": 3.6988,
      "step": 13900
    },
    {
      "epoch": 1.501453019050694,
      "grad_norm": 0.5460876822471619,
      "learning_rate": 0.0005104880939553926,
      "loss": 3.6936,
      "step": 13950
    },
    {
      "epoch": 1.5068345710903024,
      "grad_norm": 0.5820977091789246,
      "learning_rate": 0.0005101648529253313,
      "loss": 3.7014,
      "step": 14000
    },
    {
      "epoch": 1.5068345710903024,
      "eval_accuracy": 0.3535329192716649,
      "eval_loss": 3.655021905899048,
      "eval_runtime": 184.7328,
      "eval_samples_per_second": 97.498,
      "eval_steps_per_second": 6.095,
      "step": 14000
    },
    {
      "epoch": 1.5122161231299107,
      "grad_norm": 0.5606687664985657,
      "learning_rate": 0.0005098416118952699,
      "loss": 3.6949,
      "step": 14050
    },
    {
      "epoch": 1.5175976751695188,
      "grad_norm": 0.5651271343231201,
      "learning_rate": 0.0005095183708652085,
      "loss": 3.6798,
      "step": 14100
    },
    {
      "epoch": 1.5229792272091272,
      "grad_norm": 0.5150020122528076,
      "learning_rate": 0.000509195129835147,
      "loss": 3.68,
      "step": 14150
    },
    {
      "epoch": 1.5283607792487355,
      "grad_norm": 0.5977632999420166,
      "learning_rate": 0.0005088718888050856,
      "loss": 3.6851,
      "step": 14200
    },
    {
      "epoch": 1.5337423312883436,
      "grad_norm": 0.6188676953315735,
      "learning_rate": 0.0005085486477750242,
      "loss": 3.6703,
      "step": 14250
    },
    {
      "epoch": 1.5391238833279517,
      "grad_norm": 0.5410431027412415,
      "learning_rate": 0.0005082254067449629,
      "loss": 3.6897,
      "step": 14300
    },
    {
      "epoch": 1.54450543536756,
      "grad_norm": 0.5880251526832581,
      "learning_rate": 0.0005079021657149014,
      "loss": 3.7039,
      "step": 14350
    },
    {
      "epoch": 1.5498869874071683,
      "grad_norm": 0.5555596947669983,
      "learning_rate": 0.0005075789246848399,
      "loss": 3.6537,
      "step": 14400
    },
    {
      "epoch": 1.5552685394467765,
      "grad_norm": 0.5395658612251282,
      "learning_rate": 0.0005072556836547785,
      "loss": 3.6913,
      "step": 14450
    },
    {
      "epoch": 1.5606500914863846,
      "grad_norm": 0.5492722392082214,
      "learning_rate": 0.0005069324426247171,
      "loss": 3.6815,
      "step": 14500
    },
    {
      "epoch": 1.566031643525993,
      "grad_norm": 0.5226956605911255,
      "learning_rate": 0.0005066092015946557,
      "loss": 3.6833,
      "step": 14550
    },
    {
      "epoch": 1.5714131955656012,
      "grad_norm": 0.5579638481140137,
      "learning_rate": 0.0005062859605645943,
      "loss": 3.6818,
      "step": 14600
    },
    {
      "epoch": 1.5767947476052093,
      "grad_norm": 0.5651870965957642,
      "learning_rate": 0.0005059627195345329,
      "loss": 3.6777,
      "step": 14650
    },
    {
      "epoch": 1.5821762996448174,
      "grad_norm": 0.5880829095840454,
      "learning_rate": 0.0005056394785044715,
      "loss": 3.6853,
      "step": 14700
    },
    {
      "epoch": 1.5875578516844258,
      "grad_norm": 0.6156719326972961,
      "learning_rate": 0.00050531623747441,
      "loss": 3.6936,
      "step": 14750
    },
    {
      "epoch": 1.592939403724034,
      "grad_norm": 0.5112903118133545,
      "learning_rate": 0.0005049929964443486,
      "loss": 3.6819,
      "step": 14800
    },
    {
      "epoch": 1.5983209557636422,
      "grad_norm": 0.6731923818588257,
      "learning_rate": 0.0005046697554142871,
      "loss": 3.6723,
      "step": 14850
    },
    {
      "epoch": 1.6037025078032503,
      "grad_norm": 0.6107453107833862,
      "learning_rate": 0.0005043465143842258,
      "loss": 3.7043,
      "step": 14900
    },
    {
      "epoch": 1.6090840598428586,
      "grad_norm": 0.6314976811408997,
      "learning_rate": 0.0005040232733541644,
      "loss": 3.6723,
      "step": 14950
    },
    {
      "epoch": 1.614465611882467,
      "grad_norm": 0.5651878714561462,
      "learning_rate": 0.000503700032324103,
      "loss": 3.684,
      "step": 15000
    },
    {
      "epoch": 1.614465611882467,
      "eval_accuracy": 0.35533384088718994,
      "eval_loss": 3.6347434520721436,
      "eval_runtime": 184.7829,
      "eval_samples_per_second": 97.471,
      "eval_steps_per_second": 6.094,
      "step": 15000
    },
    {
      "epoch": 1.619847163922075,
      "grad_norm": 0.5857282280921936,
      "learning_rate": 0.0005033767912940415,
      "loss": 3.6826,
      "step": 15050
    },
    {
      "epoch": 1.6252287159616834,
      "grad_norm": 0.5991663336753845,
      "learning_rate": 0.0005030535502639802,
      "loss": 3.6961,
      "step": 15100
    },
    {
      "epoch": 1.6306102680012917,
      "grad_norm": 0.5337246060371399,
      "learning_rate": 0.0005027303092339187,
      "loss": 3.6949,
      "step": 15150
    },
    {
      "epoch": 1.6359918200408998,
      "grad_norm": 0.6633757948875427,
      "learning_rate": 0.0005024070682038573,
      "loss": 3.6859,
      "step": 15200
    },
    {
      "epoch": 1.641373372080508,
      "grad_norm": 0.5476294755935669,
      "learning_rate": 0.0005020838271737959,
      "loss": 3.6768,
      "step": 15250
    },
    {
      "epoch": 1.6467549241201163,
      "grad_norm": 0.5833688974380493,
      "learning_rate": 0.0005017605861437344,
      "loss": 3.6723,
      "step": 15300
    },
    {
      "epoch": 1.6521364761597246,
      "grad_norm": 0.5485004782676697,
      "learning_rate": 0.0005014373451136731,
      "loss": 3.6814,
      "step": 15350
    },
    {
      "epoch": 1.6575180281993327,
      "grad_norm": 0.5738468170166016,
      "learning_rate": 0.0005011205689042129,
      "loss": 3.6761,
      "step": 15400
    },
    {
      "epoch": 1.6628995802389408,
      "grad_norm": 0.5266857147216797,
      "learning_rate": 0.0005007973278741514,
      "loss": 3.683,
      "step": 15450
    },
    {
      "epoch": 1.6682811322785491,
      "grad_norm": 0.5718429088592529,
      "learning_rate": 0.00050047408684409,
      "loss": 3.6676,
      "step": 15500
    },
    {
      "epoch": 1.6736626843181575,
      "grad_norm": 0.5431683659553528,
      "learning_rate": 0.0005001508458140286,
      "loss": 3.6648,
      "step": 15550
    },
    {
      "epoch": 1.6790442363577656,
      "grad_norm": 0.6572660803794861,
      "learning_rate": 0.0004998276047839673,
      "loss": 3.6898,
      "step": 15600
    },
    {
      "epoch": 1.6844257883973737,
      "grad_norm": 0.5896181464195251,
      "learning_rate": 0.0004995043637539058,
      "loss": 3.6605,
      "step": 15650
    },
    {
      "epoch": 1.689807340436982,
      "grad_norm": 0.549990713596344,
      "learning_rate": 0.0004991811227238443,
      "loss": 3.657,
      "step": 15700
    },
    {
      "epoch": 1.6951888924765903,
      "grad_norm": 0.5397534966468811,
      "learning_rate": 0.0004988578816937829,
      "loss": 3.6706,
      "step": 15750
    },
    {
      "epoch": 1.7005704445161984,
      "grad_norm": 0.5501967668533325,
      "learning_rate": 0.0004985346406637215,
      "loss": 3.6697,
      "step": 15800
    },
    {
      "epoch": 1.7059519965558065,
      "grad_norm": 0.6153730154037476,
      "learning_rate": 0.0004982113996336602,
      "loss": 3.6671,
      "step": 15850
    },
    {
      "epoch": 1.7113335485954149,
      "grad_norm": 0.619310200214386,
      "learning_rate": 0.0004978881586035987,
      "loss": 3.6858,
      "step": 15900
    },
    {
      "epoch": 1.7167151006350232,
      "grad_norm": 0.5859758853912354,
      "learning_rate": 0.0004975649175735373,
      "loss": 3.6639,
      "step": 15950
    },
    {
      "epoch": 1.7220966526746313,
      "grad_norm": 0.5947201251983643,
      "learning_rate": 0.0004972416765434759,
      "loss": 3.6546,
      "step": 16000
    },
    {
      "epoch": 1.7220966526746313,
      "eval_accuracy": 0.3573106715280866,
      "eval_loss": 3.6192057132720947,
      "eval_runtime": 184.9464,
      "eval_samples_per_second": 97.385,
      "eval_steps_per_second": 6.088,
      "step": 16000
    },
    {
      "epoch": 1.7274782047142396,
      "grad_norm": 0.6014895439147949,
      "learning_rate": 0.0004969184355134145,
      "loss": 3.6555,
      "step": 16050
    },
    {
      "epoch": 1.732859756753848,
      "grad_norm": 0.6242139935493469,
      "learning_rate": 0.0004965951944833531,
      "loss": 3.6725,
      "step": 16100
    },
    {
      "epoch": 1.738241308793456,
      "grad_norm": 0.6019642949104309,
      "learning_rate": 0.0004962719534532916,
      "loss": 3.6383,
      "step": 16150
    },
    {
      "epoch": 1.7436228608330642,
      "grad_norm": 0.5532971620559692,
      "learning_rate": 0.0004959487124232302,
      "loss": 3.6586,
      "step": 16200
    },
    {
      "epoch": 1.7490044128726725,
      "grad_norm": 0.684308648109436,
      "learning_rate": 0.0004956254713931688,
      "loss": 3.6486,
      "step": 16250
    },
    {
      "epoch": 1.7543859649122808,
      "grad_norm": 0.5315883755683899,
      "learning_rate": 0.0004953022303631074,
      "loss": 3.6488,
      "step": 16300
    },
    {
      "epoch": 1.759767516951889,
      "grad_norm": 0.5260655879974365,
      "learning_rate": 0.0004949789893330459,
      "loss": 3.6647,
      "step": 16350
    },
    {
      "epoch": 1.765149068991497,
      "grad_norm": 0.5171688795089722,
      "learning_rate": 0.0004946557483029846,
      "loss": 3.6572,
      "step": 16400
    },
    {
      "epoch": 1.7705306210311054,
      "grad_norm": 0.5283517241477966,
      "learning_rate": 0.0004943325072729231,
      "loss": 3.6654,
      "step": 16450
    },
    {
      "epoch": 1.7759121730707137,
      "grad_norm": 0.5800235867500305,
      "learning_rate": 0.0004940092662428617,
      "loss": 3.673,
      "step": 16500
    },
    {
      "epoch": 1.7812937251103218,
      "grad_norm": 0.6345435380935669,
      "learning_rate": 0.0004936860252128003,
      "loss": 3.671,
      "step": 16550
    },
    {
      "epoch": 1.78667527714993,
      "grad_norm": 0.5385580658912659,
      "learning_rate": 0.0004933627841827388,
      "loss": 3.6765,
      "step": 16600
    },
    {
      "epoch": 1.7920568291895382,
      "grad_norm": 0.521317720413208,
      "learning_rate": 0.0004930395431526775,
      "loss": 3.6577,
      "step": 16650
    },
    {
      "epoch": 1.7974383812291466,
      "grad_norm": 0.596302330493927,
      "learning_rate": 0.0004927163021226161,
      "loss": 3.6477,
      "step": 16700
    },
    {
      "epoch": 1.8028199332687547,
      "grad_norm": 0.5540564656257629,
      "learning_rate": 0.0004923930610925547,
      "loss": 3.6626,
      "step": 16750
    },
    {
      "epoch": 1.8082014853083628,
      "grad_norm": 0.5950270295143127,
      "learning_rate": 0.0004920698200624932,
      "loss": 3.6484,
      "step": 16800
    },
    {
      "epoch": 1.813583037347971,
      "grad_norm": 0.5792022943496704,
      "learning_rate": 0.0004917465790324317,
      "loss": 3.6412,
      "step": 16850
    },
    {
      "epoch": 1.8189645893875794,
      "grad_norm": 0.5268118381500244,
      "learning_rate": 0.0004914233380023704,
      "loss": 3.6561,
      "step": 16900
    },
    {
      "epoch": 1.8243461414271875,
      "grad_norm": 0.5259914994239807,
      "learning_rate": 0.0004911000969723089,
      "loss": 3.6361,
      "step": 16950
    },
    {
      "epoch": 1.8297276934667959,
      "grad_norm": 0.5924632549285889,
      "learning_rate": 0.0004907768559422476,
      "loss": 3.6475,
      "step": 17000
    },
    {
      "epoch": 1.8297276934667959,
      "eval_accuracy": 0.3591788492811688,
      "eval_loss": 3.601672410964966,
      "eval_runtime": 184.7368,
      "eval_samples_per_second": 97.495,
      "eval_steps_per_second": 6.095,
      "step": 17000
    },
    {
      "epoch": 1.8351092455064042,
      "grad_norm": 0.5890408754348755,
      "learning_rate": 0.0004904536149121861,
      "loss": 3.6432,
      "step": 17050
    },
    {
      "epoch": 1.8404907975460123,
      "grad_norm": 0.5994479656219482,
      "learning_rate": 0.0004901303738821248,
      "loss": 3.645,
      "step": 17100
    },
    {
      "epoch": 1.8458723495856204,
      "grad_norm": 0.5794470310211182,
      "learning_rate": 0.0004898071328520633,
      "loss": 3.6375,
      "step": 17150
    },
    {
      "epoch": 1.8512539016252287,
      "grad_norm": 0.5518653392791748,
      "learning_rate": 0.0004894838918220019,
      "loss": 3.6343,
      "step": 17200
    },
    {
      "epoch": 1.856635453664837,
      "grad_norm": 0.5658548474311829,
      "learning_rate": 0.0004891606507919405,
      "loss": 3.6607,
      "step": 17250
    },
    {
      "epoch": 1.8620170057044452,
      "grad_norm": 0.5588201880455017,
      "learning_rate": 0.000488837409761879,
      "loss": 3.6551,
      "step": 17300
    },
    {
      "epoch": 1.8673985577440533,
      "grad_norm": 0.5978080034255981,
      "learning_rate": 0.0004885141687318177,
      "loss": 3.6435,
      "step": 17350
    },
    {
      "epoch": 1.8727801097836616,
      "grad_norm": 0.5438842177391052,
      "learning_rate": 0.00048819092770175623,
      "loss": 3.6491,
      "step": 17400
    },
    {
      "epoch": 1.87816166182327,
      "grad_norm": 0.5227386951446533,
      "learning_rate": 0.00048787415149229604,
      "loss": 3.6431,
      "step": 17450
    },
    {
      "epoch": 1.883543213862878,
      "grad_norm": 0.5770353078842163,
      "learning_rate": 0.00048755091046223464,
      "loss": 3.652,
      "step": 17500
    },
    {
      "epoch": 1.8889247659024861,
      "grad_norm": 0.6320988535881042,
      "learning_rate": 0.0004872276694321732,
      "loss": 3.6296,
      "step": 17550
    },
    {
      "epoch": 1.8943063179420945,
      "grad_norm": 0.5457765460014343,
      "learning_rate": 0.00048690442840211177,
      "loss": 3.6411,
      "step": 17600
    },
    {
      "epoch": 1.8996878699817028,
      "grad_norm": 0.5683311820030212,
      "learning_rate": 0.0004865811873720504,
      "loss": 3.6369,
      "step": 17650
    },
    {
      "epoch": 1.905069422021311,
      "grad_norm": 0.6170719861984253,
      "learning_rate": 0.00048625794634198896,
      "loss": 3.6392,
      "step": 17700
    },
    {
      "epoch": 1.910450974060919,
      "grad_norm": 0.5767971277236938,
      "learning_rate": 0.00048593470531192756,
      "loss": 3.6482,
      "step": 17750
    },
    {
      "epoch": 1.9158325261005273,
      "grad_norm": 0.5953500270843506,
      "learning_rate": 0.00048561146428186615,
      "loss": 3.6474,
      "step": 17800
    },
    {
      "epoch": 1.9212140781401357,
      "grad_norm": 0.5317740440368652,
      "learning_rate": 0.0004852882232518047,
      "loss": 3.6312,
      "step": 17850
    },
    {
      "epoch": 1.9265956301797438,
      "grad_norm": 0.5526648163795471,
      "learning_rate": 0.00048496498222174334,
      "loss": 3.6434,
      "step": 17900
    },
    {
      "epoch": 1.931977182219352,
      "grad_norm": 0.5383221507072449,
      "learning_rate": 0.00048464174119168193,
      "loss": 3.6373,
      "step": 17950
    },
    {
      "epoch": 1.9373587342589604,
      "grad_norm": 0.567500114440918,
      "learning_rate": 0.0004843185001616205,
      "loss": 3.6584,
      "step": 18000
    },
    {
      "epoch": 1.9373587342589604,
      "eval_accuracy": 0.3605316863873469,
      "eval_loss": 3.5917465686798096,
      "eval_runtime": 184.8222,
      "eval_samples_per_second": 97.45,
      "eval_steps_per_second": 6.092,
      "step": 18000
    },
    {
      "epoch": 1.9427402862985685,
      "grad_norm": 0.5618447065353394,
      "learning_rate": 0.00048399525913155907,
      "loss": 3.6331,
      "step": 18050
    },
    {
      "epoch": 1.9481218383381766,
      "grad_norm": 0.5482614636421204,
      "learning_rate": 0.0004836720181014976,
      "loss": 3.6441,
      "step": 18100
    },
    {
      "epoch": 1.953503390377785,
      "grad_norm": 0.5669938325881958,
      "learning_rate": 0.0004833487770714362,
      "loss": 3.6514,
      "step": 18150
    },
    {
      "epoch": 1.9588849424173933,
      "grad_norm": 0.5920496582984924,
      "learning_rate": 0.00048302553604137485,
      "loss": 3.6434,
      "step": 18200
    },
    {
      "epoch": 1.9642664944570014,
      "grad_norm": 0.5437381267547607,
      "learning_rate": 0.0004827022950113134,
      "loss": 3.664,
      "step": 18250
    },
    {
      "epoch": 1.9696480464966095,
      "grad_norm": 0.5581496357917786,
      "learning_rate": 0.000482379053981252,
      "loss": 3.6224,
      "step": 18300
    },
    {
      "epoch": 1.9750295985362178,
      "grad_norm": 0.5657931566238403,
      "learning_rate": 0.0004820558129511906,
      "loss": 3.6262,
      "step": 18350
    },
    {
      "epoch": 1.9804111505758262,
      "grad_norm": 0.5733588933944702,
      "learning_rate": 0.0004817325719211291,
      "loss": 3.6271,
      "step": 18400
    },
    {
      "epoch": 1.9857927026154343,
      "grad_norm": 0.5454577803611755,
      "learning_rate": 0.0004814093308910677,
      "loss": 3.6385,
      "step": 18450
    },
    {
      "epoch": 1.9911742546550424,
      "grad_norm": 0.5400912165641785,
      "learning_rate": 0.00048108608986100637,
      "loss": 3.6269,
      "step": 18500
    },
    {
      "epoch": 1.9965558066946507,
      "grad_norm": 0.5676894187927246,
      "learning_rate": 0.0004807628488309449,
      "loss": 3.6297,
      "step": 18550
    },
    {
      "epoch": 2.001937358734259,
      "grad_norm": 0.5944864153862,
      "learning_rate": 0.0004804396078008835,
      "loss": 3.578,
      "step": 18600
    },
    {
      "epoch": 2.007318910773867,
      "grad_norm": 0.5926184058189392,
      "learning_rate": 0.00048011636677082204,
      "loss": 3.5433,
      "step": 18650
    },
    {
      "epoch": 2.0127004628134753,
      "grad_norm": 0.507103443145752,
      "learning_rate": 0.00047979312574076064,
      "loss": 3.5372,
      "step": 18700
    },
    {
      "epoch": 2.018082014853084,
      "grad_norm": 0.5739589929580688,
      "learning_rate": 0.0004794698847106992,
      "loss": 3.5428,
      "step": 18750
    },
    {
      "epoch": 2.023463566892692,
      "grad_norm": 0.5899766087532043,
      "learning_rate": 0.0004791466436806378,
      "loss": 3.5556,
      "step": 18800
    },
    {
      "epoch": 2.0288451189323,
      "grad_norm": 0.603664755821228,
      "learning_rate": 0.0004788234026505764,
      "loss": 3.5476,
      "step": 18850
    },
    {
      "epoch": 2.034226670971908,
      "grad_norm": 0.6328676342964172,
      "learning_rate": 0.00047850016162051496,
      "loss": 3.5374,
      "step": 18900
    },
    {
      "epoch": 2.0396082230115167,
      "grad_norm": 0.5715523362159729,
      "learning_rate": 0.00047817692059045356,
      "loss": 3.5504,
      "step": 18950
    },
    {
      "epoch": 2.044989775051125,
      "grad_norm": 0.554076075553894,
      "learning_rate": 0.00047785367956039215,
      "loss": 3.5579,
      "step": 19000
    },
    {
      "epoch": 2.044989775051125,
      "eval_accuracy": 0.3616910127003274,
      "eval_loss": 3.57794189453125,
      "eval_runtime": 184.5487,
      "eval_samples_per_second": 97.595,
      "eval_steps_per_second": 6.101,
      "step": 19000
    },
    {
      "epoch": 2.050371327090733,
      "grad_norm": 0.5813206434249878,
      "learning_rate": 0.00047753043853033075,
      "loss": 3.5526,
      "step": 19050
    },
    {
      "epoch": 2.055752879130341,
      "grad_norm": 0.5949262976646423,
      "learning_rate": 0.00047720719750026934,
      "loss": 3.5472,
      "step": 19100
    },
    {
      "epoch": 2.0611344311699495,
      "grad_norm": 0.5201438069343567,
      "learning_rate": 0.00047688395647020793,
      "loss": 3.5551,
      "step": 19150
    },
    {
      "epoch": 2.0665159832095576,
      "grad_norm": 0.5755208730697632,
      "learning_rate": 0.0004765607154401465,
      "loss": 3.5279,
      "step": 19200
    },
    {
      "epoch": 2.0718975352491658,
      "grad_norm": 0.5886051654815674,
      "learning_rate": 0.00047623747441008507,
      "loss": 3.5456,
      "step": 19250
    },
    {
      "epoch": 2.0772790872887743,
      "grad_norm": 0.581571102142334,
      "learning_rate": 0.0004759142333800236,
      "loss": 3.5523,
      "step": 19300
    },
    {
      "epoch": 2.0826606393283824,
      "grad_norm": 0.5398459434509277,
      "learning_rate": 0.00047559099234996226,
      "loss": 3.5488,
      "step": 19350
    },
    {
      "epoch": 2.0880421913679905,
      "grad_norm": 0.5748805403709412,
      "learning_rate": 0.00047526775131990085,
      "loss": 3.5538,
      "step": 19400
    },
    {
      "epoch": 2.0934237434075986,
      "grad_norm": 0.5900923013687134,
      "learning_rate": 0.0004749445102898394,
      "loss": 3.5743,
      "step": 19450
    },
    {
      "epoch": 2.098805295447207,
      "grad_norm": 0.5444471836090088,
      "learning_rate": 0.000474621269259778,
      "loss": 3.5705,
      "step": 19500
    },
    {
      "epoch": 2.1041868474868153,
      "grad_norm": 0.5641753673553467,
      "learning_rate": 0.0004742980282297166,
      "loss": 3.5555,
      "step": 19550
    },
    {
      "epoch": 2.1095683995264234,
      "grad_norm": 0.5689592957496643,
      "learning_rate": 0.0004739812520202564,
      "loss": 3.5609,
      "step": 19600
    },
    {
      "epoch": 2.1149499515660315,
      "grad_norm": 0.6442001461982727,
      "learning_rate": 0.000473658010990195,
      "loss": 3.5649,
      "step": 19650
    },
    {
      "epoch": 2.12033150360564,
      "grad_norm": 0.5538535118103027,
      "learning_rate": 0.0004733347699601336,
      "loss": 3.5624,
      "step": 19700
    },
    {
      "epoch": 2.125713055645248,
      "grad_norm": 0.5702688097953796,
      "learning_rate": 0.0004730115289300722,
      "loss": 3.5425,
      "step": 19750
    },
    {
      "epoch": 2.1310946076848563,
      "grad_norm": 0.5623447895050049,
      "learning_rate": 0.000472694752720612,
      "loss": 3.5462,
      "step": 19800
    },
    {
      "epoch": 2.1364761597244644,
      "grad_norm": 0.6005106568336487,
      "learning_rate": 0.0004723715116905506,
      "loss": 3.5623,
      "step": 19850
    },
    {
      "epoch": 2.141857711764073,
      "grad_norm": 0.530021607875824,
      "learning_rate": 0.0004720482706604891,
      "loss": 3.552,
      "step": 19900
    },
    {
      "epoch": 2.147239263803681,
      "grad_norm": 0.6486440896987915,
      "learning_rate": 0.0004717250296304277,
      "loss": 3.5587,
      "step": 19950
    },
    {
      "epoch": 2.152620815843289,
      "grad_norm": 0.5921491980552673,
      "learning_rate": 0.0004714017886003663,
      "loss": 3.5664,
      "step": 20000
    },
    {
      "epoch": 2.152620815843289,
      "eval_accuracy": 0.36275624561246056,
      "eval_loss": 3.568859815597534,
      "eval_runtime": 184.9316,
      "eval_samples_per_second": 97.393,
      "eval_steps_per_second": 6.089,
      "step": 20000
    },
    {
      "epoch": 2.1580023678828972,
      "grad_norm": 0.5587154030799866,
      "learning_rate": 0.00047107854757030485,
      "loss": 3.5344,
      "step": 20050
    },
    {
      "epoch": 2.163383919922506,
      "grad_norm": 0.6131019592285156,
      "learning_rate": 0.0004707553065402435,
      "loss": 3.563,
      "step": 20100
    },
    {
      "epoch": 2.168765471962114,
      "grad_norm": 0.602177083492279,
      "learning_rate": 0.0004704320655101821,
      "loss": 3.5386,
      "step": 20150
    },
    {
      "epoch": 2.174147024001722,
      "grad_norm": 0.5748916864395142,
      "learning_rate": 0.00047010882448012063,
      "loss": 3.5412,
      "step": 20200
    },
    {
      "epoch": 2.1795285760413305,
      "grad_norm": 0.5983401536941528,
      "learning_rate": 0.0004697855834500592,
      "loss": 3.5439,
      "step": 20250
    },
    {
      "epoch": 2.1849101280809387,
      "grad_norm": 0.5558173656463623,
      "learning_rate": 0.00046946234241999776,
      "loss": 3.5476,
      "step": 20300
    },
    {
      "epoch": 2.1902916801205468,
      "grad_norm": 0.5931748151779175,
      "learning_rate": 0.00046913910138993636,
      "loss": 3.5686,
      "step": 20350
    },
    {
      "epoch": 2.195673232160155,
      "grad_norm": 0.5662775635719299,
      "learning_rate": 0.000468815860359875,
      "loss": 3.5691,
      "step": 20400
    },
    {
      "epoch": 2.2010547841997634,
      "grad_norm": 0.6415780186653137,
      "learning_rate": 0.00046849261932981355,
      "loss": 3.54,
      "step": 20450
    },
    {
      "epoch": 2.2064363362393715,
      "grad_norm": 0.5790057182312012,
      "learning_rate": 0.00046816937829975214,
      "loss": 3.5471,
      "step": 20500
    },
    {
      "epoch": 2.2118178882789796,
      "grad_norm": 0.559398353099823,
      "learning_rate": 0.00046784613726969074,
      "loss": 3.5688,
      "step": 20550
    },
    {
      "epoch": 2.2171994403185877,
      "grad_norm": 0.5323479175567627,
      "learning_rate": 0.0004675228962396293,
      "loss": 3.5536,
      "step": 20600
    },
    {
      "epoch": 2.2225809923581963,
      "grad_norm": 0.5640736818313599,
      "learning_rate": 0.0004671996552095679,
      "loss": 3.5444,
      "step": 20650
    },
    {
      "epoch": 2.2279625443978044,
      "grad_norm": 0.6212958097457886,
      "learning_rate": 0.0004668764141795065,
      "loss": 3.5705,
      "step": 20700
    },
    {
      "epoch": 2.2333440964374125,
      "grad_norm": 0.5601683855056763,
      "learning_rate": 0.00046655317314944506,
      "loss": 3.5538,
      "step": 20750
    },
    {
      "epoch": 2.2387256484770206,
      "grad_norm": 0.5923741459846497,
      "learning_rate": 0.00046622993211938366,
      "loss": 3.5587,
      "step": 20800
    },
    {
      "epoch": 2.244107200516629,
      "grad_norm": 0.5725792646408081,
      "learning_rate": 0.0004659066910893222,
      "loss": 3.5558,
      "step": 20850
    },
    {
      "epoch": 2.2494887525562373,
      "grad_norm": 0.577241063117981,
      "learning_rate": 0.0004655834500592608,
      "loss": 3.5422,
      "step": 20900
    },
    {
      "epoch": 2.2548703045958454,
      "grad_norm": 0.5885173082351685,
      "learning_rate": 0.00046526020902919944,
      "loss": 3.5613,
      "step": 20950
    },
    {
      "epoch": 2.2602518566354535,
      "grad_norm": 0.6301099061965942,
      "learning_rate": 0.000464936967999138,
      "loss": 3.5668,
      "step": 21000
    },
    {
      "epoch": 2.2602518566354535,
      "eval_accuracy": 0.3639710935511143,
      "eval_loss": 3.557680606842041,
      "eval_runtime": 184.4972,
      "eval_samples_per_second": 97.622,
      "eval_steps_per_second": 6.103,
      "step": 21000
    },
    {
      "epoch": 2.265633408675062,
      "grad_norm": 0.6086539030075073,
      "learning_rate": 0.0004646137269690766,
      "loss": 3.5513,
      "step": 21050
    },
    {
      "epoch": 2.27101496071467,
      "grad_norm": 0.5513246059417725,
      "learning_rate": 0.00046429048593901517,
      "loss": 3.5549,
      "step": 21100
    },
    {
      "epoch": 2.2763965127542782,
      "grad_norm": 0.5785967111587524,
      "learning_rate": 0.0004639672449089537,
      "loss": 3.5596,
      "step": 21150
    },
    {
      "epoch": 2.281778064793887,
      "grad_norm": 0.5956328511238098,
      "learning_rate": 0.0004636440038788923,
      "loss": 3.5427,
      "step": 21200
    },
    {
      "epoch": 2.287159616833495,
      "grad_norm": 0.6081483364105225,
      "learning_rate": 0.00046332076284883095,
      "loss": 3.5325,
      "step": 21250
    },
    {
      "epoch": 2.292541168873103,
      "grad_norm": 0.5518943071365356,
      "learning_rate": 0.0004629975218187695,
      "loss": 3.5368,
      "step": 21300
    },
    {
      "epoch": 2.297922720912711,
      "grad_norm": 0.5766976475715637,
      "learning_rate": 0.0004626742807887081,
      "loss": 3.5642,
      "step": 21350
    },
    {
      "epoch": 2.303304272952319,
      "grad_norm": 0.582249104976654,
      "learning_rate": 0.00046235103975864663,
      "loss": 3.5492,
      "step": 21400
    },
    {
      "epoch": 2.3086858249919278,
      "grad_norm": 0.6385564208030701,
      "learning_rate": 0.0004620277987285852,
      "loss": 3.5574,
      "step": 21450
    },
    {
      "epoch": 2.314067377031536,
      "grad_norm": 0.5533836483955383,
      "learning_rate": 0.0004617045576985239,
      "loss": 3.5461,
      "step": 21500
    },
    {
      "epoch": 2.319448929071144,
      "grad_norm": 0.6168516278266907,
      "learning_rate": 0.0004613813166684624,
      "loss": 3.536,
      "step": 21550
    },
    {
      "epoch": 2.3248304811107525,
      "grad_norm": 0.5422675013542175,
      "learning_rate": 0.000461058075638401,
      "loss": 3.5478,
      "step": 21600
    },
    {
      "epoch": 2.3302120331503606,
      "grad_norm": 0.673985481262207,
      "learning_rate": 0.00046073483460833955,
      "loss": 3.5648,
      "step": 21650
    },
    {
      "epoch": 2.3355935851899687,
      "grad_norm": 0.6507932543754578,
      "learning_rate": 0.00046041159357827814,
      "loss": 3.5514,
      "step": 21700
    },
    {
      "epoch": 2.340975137229577,
      "grad_norm": 0.5993643403053284,
      "learning_rate": 0.00046008835254821674,
      "loss": 3.5723,
      "step": 21750
    },
    {
      "epoch": 2.3463566892691854,
      "grad_norm": 0.629294753074646,
      "learning_rate": 0.0004597651115181554,
      "loss": 3.5385,
      "step": 21800
    },
    {
      "epoch": 2.3517382413087935,
      "grad_norm": 0.5397011041641235,
      "learning_rate": 0.00045944187048809393,
      "loss": 3.5378,
      "step": 21850
    },
    {
      "epoch": 2.3571197933484016,
      "grad_norm": 0.5139660835266113,
      "learning_rate": 0.0004591186294580325,
      "loss": 3.5584,
      "step": 21900
    },
    {
      "epoch": 2.3625013453880097,
      "grad_norm": 0.6737929582595825,
      "learning_rate": 0.00045879538842797106,
      "loss": 3.5637,
      "step": 21950
    },
    {
      "epoch": 2.3678828974276183,
      "grad_norm": 0.5924611687660217,
      "learning_rate": 0.00045847214739790966,
      "loss": 3.5642,
      "step": 22000
    },
    {
      "epoch": 2.3678828974276183,
      "eval_accuracy": 0.36480739482862234,
      "eval_loss": 3.550454616546631,
      "eval_runtime": 184.9625,
      "eval_samples_per_second": 97.376,
      "eval_steps_per_second": 6.088,
      "step": 22000
    },
    {
      "epoch": 2.3732644494672264,
      "grad_norm": 0.5367729067802429,
      "learning_rate": 0.0004581489063678482,
      "loss": 3.5564,
      "step": 22050
    },
    {
      "epoch": 2.3786460015068345,
      "grad_norm": 0.6676405072212219,
      "learning_rate": 0.00045782566533778685,
      "loss": 3.5521,
      "step": 22100
    },
    {
      "epoch": 2.384027553546443,
      "grad_norm": 0.5662420988082886,
      "learning_rate": 0.00045750242430772544,
      "loss": 3.5702,
      "step": 22150
    },
    {
      "epoch": 2.389409105586051,
      "grad_norm": 0.5759710073471069,
      "learning_rate": 0.000457179183277664,
      "loss": 3.5667,
      "step": 22200
    },
    {
      "epoch": 2.3947906576256592,
      "grad_norm": 0.5874856114387512,
      "learning_rate": 0.0004568559422476026,
      "loss": 3.5669,
      "step": 22250
    },
    {
      "epoch": 2.4001722096652673,
      "grad_norm": 0.6054186820983887,
      "learning_rate": 0.00045653270121754117,
      "loss": 3.5455,
      "step": 22300
    },
    {
      "epoch": 2.4055537617048754,
      "grad_norm": 0.5471639037132263,
      "learning_rate": 0.0004562094601874797,
      "loss": 3.5433,
      "step": 22350
    },
    {
      "epoch": 2.410935313744484,
      "grad_norm": 0.5708964467048645,
      "learning_rate": 0.00045588621915741836,
      "loss": 3.5434,
      "step": 22400
    },
    {
      "epoch": 2.416316865784092,
      "grad_norm": 0.5709865093231201,
      "learning_rate": 0.00045556297812735696,
      "loss": 3.5641,
      "step": 22450
    },
    {
      "epoch": 2.4216984178237,
      "grad_norm": 0.608943521976471,
      "learning_rate": 0.0004552397370972955,
      "loss": 3.5373,
      "step": 22500
    },
    {
      "epoch": 2.4270799698633088,
      "grad_norm": 0.5688168406486511,
      "learning_rate": 0.0004549164960672341,
      "loss": 3.5529,
      "step": 22550
    },
    {
      "epoch": 2.432461521902917,
      "grad_norm": 0.5552303791046143,
      "learning_rate": 0.00045459325503717263,
      "loss": 3.5403,
      "step": 22600
    },
    {
      "epoch": 2.437843073942525,
      "grad_norm": 0.5680282711982727,
      "learning_rate": 0.0004542700140071113,
      "loss": 3.5457,
      "step": 22650
    },
    {
      "epoch": 2.443224625982133,
      "grad_norm": 0.584951639175415,
      "learning_rate": 0.0004539467729770499,
      "loss": 3.5575,
      "step": 22700
    },
    {
      "epoch": 2.4486061780217416,
      "grad_norm": 0.5572960376739502,
      "learning_rate": 0.0004536235319469884,
      "loss": 3.5445,
      "step": 22750
    },
    {
      "epoch": 2.4539877300613497,
      "grad_norm": 0.5390388369560242,
      "learning_rate": 0.000453300290916927,
      "loss": 3.5575,
      "step": 22800
    },
    {
      "epoch": 2.459369282100958,
      "grad_norm": 0.6123374700546265,
      "learning_rate": 0.0004529770498868656,
      "loss": 3.5767,
      "step": 22850
    },
    {
      "epoch": 2.464750834140566,
      "grad_norm": 0.5894783139228821,
      "learning_rate": 0.00045265380885680414,
      "loss": 3.5542,
      "step": 22900
    },
    {
      "epoch": 2.4701323861801745,
      "grad_norm": 0.5763325095176697,
      "learning_rate": 0.0004523305678267428,
      "loss": 3.5624,
      "step": 22950
    },
    {
      "epoch": 2.4755139382197826,
      "grad_norm": 0.5496013760566711,
      "learning_rate": 0.0004520073267966814,
      "loss": 3.5787,
      "step": 23000
    },
    {
      "epoch": 2.4755139382197826,
      "eval_accuracy": 0.3661909807020519,
      "eval_loss": 3.5384585857391357,
      "eval_runtime": 184.7612,
      "eval_samples_per_second": 97.483,
      "eval_steps_per_second": 6.094,
      "step": 23000
    },
    {
      "epoch": 2.4808954902593907,
      "grad_norm": 0.6314433217048645,
      "learning_rate": 0.00045168408576661993,
      "loss": 3.538,
      "step": 23050
    },
    {
      "epoch": 2.4862770422989993,
      "grad_norm": 0.5926424860954285,
      "learning_rate": 0.0004513608447365585,
      "loss": 3.5478,
      "step": 23100
    },
    {
      "epoch": 2.4916585943386074,
      "grad_norm": 0.5751678943634033,
      "learning_rate": 0.00045103760370649706,
      "loss": 3.5523,
      "step": 23150
    },
    {
      "epoch": 2.4970401463782155,
      "grad_norm": 0.5919656753540039,
      "learning_rate": 0.00045071436267643566,
      "loss": 3.5424,
      "step": 23200
    },
    {
      "epoch": 2.5024216984178236,
      "grad_norm": 0.5607296228408813,
      "learning_rate": 0.0004503911216463743,
      "loss": 3.5337,
      "step": 23250
    },
    {
      "epoch": 2.5078032504574317,
      "grad_norm": 0.5334439873695374,
      "learning_rate": 0.00045006788061631285,
      "loss": 3.5375,
      "step": 23300
    },
    {
      "epoch": 2.5131848024970402,
      "grad_norm": 0.5940549969673157,
      "learning_rate": 0.00044974463958625144,
      "loss": 3.5429,
      "step": 23350
    },
    {
      "epoch": 2.5185663545366483,
      "grad_norm": 0.5540521740913391,
      "learning_rate": 0.00044942139855619004,
      "loss": 3.5407,
      "step": 23400
    },
    {
      "epoch": 2.5239479065762565,
      "grad_norm": 0.6090623140335083,
      "learning_rate": 0.0004490981575261286,
      "loss": 3.5343,
      "step": 23450
    },
    {
      "epoch": 2.529329458615865,
      "grad_norm": 0.5577390789985657,
      "learning_rate": 0.0004487749164960672,
      "loss": 3.5486,
      "step": 23500
    },
    {
      "epoch": 2.534711010655473,
      "grad_norm": 0.5548477172851562,
      "learning_rate": 0.0004484516754660058,
      "loss": 3.5276,
      "step": 23550
    },
    {
      "epoch": 2.540092562695081,
      "grad_norm": 0.6063026189804077,
      "learning_rate": 0.00044812843443594436,
      "loss": 3.5385,
      "step": 23600
    },
    {
      "epoch": 2.5454741147346893,
      "grad_norm": 0.5368787050247192,
      "learning_rate": 0.00044780519340588296,
      "loss": 3.5378,
      "step": 23650
    },
    {
      "epoch": 2.550855666774298,
      "grad_norm": 0.6063023805618286,
      "learning_rate": 0.0004474819523758215,
      "loss": 3.5421,
      "step": 23700
    },
    {
      "epoch": 2.556237218813906,
      "grad_norm": 0.5546424984931946,
      "learning_rate": 0.0004471587113457601,
      "loss": 3.5394,
      "step": 23750
    },
    {
      "epoch": 2.561618770853514,
      "grad_norm": 0.5690875053405762,
      "learning_rate": 0.00044683547031569874,
      "loss": 3.5454,
      "step": 23800
    },
    {
      "epoch": 2.567000322893122,
      "grad_norm": 0.5386408567428589,
      "learning_rate": 0.0004465186941062385,
      "loss": 3.5472,
      "step": 23850
    },
    {
      "epoch": 2.5723818749327307,
      "grad_norm": 0.5763099193572998,
      "learning_rate": 0.00044619545307617714,
      "loss": 3.5556,
      "step": 23900
    },
    {
      "epoch": 2.577763426972339,
      "grad_norm": 0.5599791407585144,
      "learning_rate": 0.0004458722120461157,
      "loss": 3.5531,
      "step": 23950
    },
    {
      "epoch": 2.583144979011947,
      "grad_norm": 0.5776103138923645,
      "learning_rate": 0.0004455489710160543,
      "loss": 3.5471,
      "step": 24000
    },
    {
      "epoch": 2.583144979011947,
      "eval_accuracy": 0.3669588306602369,
      "eval_loss": 3.5304083824157715,
      "eval_runtime": 185.0981,
      "eval_samples_per_second": 97.305,
      "eval_steps_per_second": 6.083,
      "step": 24000
    },
    {
      "epoch": 2.5885265310515555,
      "grad_norm": 0.5959745049476624,
      "learning_rate": 0.0004452257299859928,
      "loss": 3.5302,
      "step": 24050
    },
    {
      "epoch": 2.5939080830911636,
      "grad_norm": 0.5823758840560913,
      "learning_rate": 0.0004449024889559314,
      "loss": 3.5519,
      "step": 24100
    },
    {
      "epoch": 2.5992896351307717,
      "grad_norm": 0.5374528765678406,
      "learning_rate": 0.00044457924792587,
      "loss": 3.541,
      "step": 24150
    },
    {
      "epoch": 2.60467118717038,
      "grad_norm": 0.5732862949371338,
      "learning_rate": 0.0004442560068958086,
      "loss": 3.5239,
      "step": 24200
    },
    {
      "epoch": 2.610052739209988,
      "grad_norm": 0.5892502069473267,
      "learning_rate": 0.0004439327658657472,
      "loss": 3.5252,
      "step": 24250
    },
    {
      "epoch": 2.6154342912495965,
      "grad_norm": 0.572582483291626,
      "learning_rate": 0.0004436095248356858,
      "loss": 3.5351,
      "step": 24300
    },
    {
      "epoch": 2.6208158432892046,
      "grad_norm": 0.5686654448509216,
      "learning_rate": 0.00044328628380562433,
      "loss": 3.5611,
      "step": 24350
    },
    {
      "epoch": 2.6261973953288127,
      "grad_norm": 0.6537500619888306,
      "learning_rate": 0.0004429630427755629,
      "loss": 3.5398,
      "step": 24400
    },
    {
      "epoch": 2.6315789473684212,
      "grad_norm": 0.5685673356056213,
      "learning_rate": 0.0004426398017455016,
      "loss": 3.5416,
      "step": 24450
    },
    {
      "epoch": 2.6369604994080293,
      "grad_norm": 0.5964141488075256,
      "learning_rate": 0.0004423165607154401,
      "loss": 3.5345,
      "step": 24500
    },
    {
      "epoch": 2.6423420514476375,
      "grad_norm": 0.5742844343185425,
      "learning_rate": 0.0004419933196853787,
      "loss": 3.5365,
      "step": 24550
    },
    {
      "epoch": 2.6477236034872456,
      "grad_norm": 0.6200198531150818,
      "learning_rate": 0.00044167007865531725,
      "loss": 3.5483,
      "step": 24600
    },
    {
      "epoch": 2.653105155526854,
      "grad_norm": 0.5747848153114319,
      "learning_rate": 0.00044134683762525584,
      "loss": 3.5481,
      "step": 24650
    },
    {
      "epoch": 2.658486707566462,
      "grad_norm": 0.5651057958602905,
      "learning_rate": 0.00044102359659519444,
      "loss": 3.5375,
      "step": 24700
    },
    {
      "epoch": 2.6638682596060703,
      "grad_norm": 0.6583673357963562,
      "learning_rate": 0.00044070035556513303,
      "loss": 3.5528,
      "step": 24750
    },
    {
      "epoch": 2.6692498116456784,
      "grad_norm": 0.6136596202850342,
      "learning_rate": 0.00044037711453507163,
      "loss": 3.5511,
      "step": 24800
    },
    {
      "epoch": 2.674631363685287,
      "grad_norm": 0.534138560295105,
      "learning_rate": 0.0004400538735050102,
      "loss": 3.5538,
      "step": 24850
    },
    {
      "epoch": 2.680012915724895,
      "grad_norm": 0.5683159828186035,
      "learning_rate": 0.00043973063247494876,
      "loss": 3.5404,
      "step": 24900
    },
    {
      "epoch": 2.685394467764503,
      "grad_norm": 0.5965563058853149,
      "learning_rate": 0.00043940739144488736,
      "loss": 3.5361,
      "step": 24950
    },
    {
      "epoch": 2.6907760198041117,
      "grad_norm": 0.6091514825820923,
      "learning_rate": 0.0004390841504148259,
      "loss": 3.5407,
      "step": 25000
    },
    {
      "epoch": 2.6907760198041117,
      "eval_accuracy": 0.368015914605784,
      "eval_loss": 3.5202255249023438,
      "eval_runtime": 184.4748,
      "eval_samples_per_second": 97.634,
      "eval_steps_per_second": 6.104,
      "step": 25000
    },
    {
      "epoch": 2.69615757184372,
      "grad_norm": 0.5996729135513306,
      "learning_rate": 0.00043876090938476455,
      "loss": 3.5246,
      "step": 25050
    },
    {
      "epoch": 2.701539123883328,
      "grad_norm": 0.5679774284362793,
      "learning_rate": 0.00043843766835470314,
      "loss": 3.5314,
      "step": 25100
    },
    {
      "epoch": 2.706920675922936,
      "grad_norm": 0.6458651423454285,
      "learning_rate": 0.0004381144273246417,
      "loss": 3.5488,
      "step": 25150
    },
    {
      "epoch": 2.712302227962544,
      "grad_norm": 0.5411641597747803,
      "learning_rate": 0.0004377911862945803,
      "loss": 3.5364,
      "step": 25200
    },
    {
      "epoch": 2.7176837800021527,
      "grad_norm": 0.5480161309242249,
      "learning_rate": 0.00043746794526451887,
      "loss": 3.5359,
      "step": 25250
    },
    {
      "epoch": 2.723065332041761,
      "grad_norm": 0.5879181623458862,
      "learning_rate": 0.00043714470423445747,
      "loss": 3.5329,
      "step": 25300
    },
    {
      "epoch": 2.728446884081369,
      "grad_norm": 0.5946080088615417,
      "learning_rate": 0.00043682146320439606,
      "loss": 3.54,
      "step": 25350
    },
    {
      "epoch": 2.7338284361209775,
      "grad_norm": 0.5855832099914551,
      "learning_rate": 0.00043649822217433466,
      "loss": 3.5489,
      "step": 25400
    },
    {
      "epoch": 2.7392099881605856,
      "grad_norm": 0.5638436675071716,
      "learning_rate": 0.0004361749811442732,
      "loss": 3.5306,
      "step": 25450
    },
    {
      "epoch": 2.7445915402001937,
      "grad_norm": 0.5559232831001282,
      "learning_rate": 0.0004358517401142118,
      "loss": 3.5255,
      "step": 25500
    },
    {
      "epoch": 2.749973092239802,
      "grad_norm": 0.5973818302154541,
      "learning_rate": 0.00043552849908415033,
      "loss": 3.5437,
      "step": 25550
    },
    {
      "epoch": 2.7553546442794103,
      "grad_norm": 0.5868799090385437,
      "learning_rate": 0.000435205258054089,
      "loss": 3.5466,
      "step": 25600
    },
    {
      "epoch": 2.7607361963190185,
      "grad_norm": 0.6294220089912415,
      "learning_rate": 0.0004348820170240276,
      "loss": 3.543,
      "step": 25650
    },
    {
      "epoch": 2.7661177483586266,
      "grad_norm": 0.5231223702430725,
      "learning_rate": 0.0004345587759939661,
      "loss": 3.5294,
      "step": 25700
    },
    {
      "epoch": 2.7714993003982347,
      "grad_norm": 0.634873628616333,
      "learning_rate": 0.0004342355349639047,
      "loss": 3.537,
      "step": 25750
    },
    {
      "epoch": 2.776880852437843,
      "grad_norm": 0.5652748346328735,
      "learning_rate": 0.0004339122939338433,
      "loss": 3.5232,
      "step": 25800
    },
    {
      "epoch": 2.7822624044774513,
      "grad_norm": 0.6224278807640076,
      "learning_rate": 0.00043358905290378184,
      "loss": 3.5462,
      "step": 25850
    },
    {
      "epoch": 2.7876439565170594,
      "grad_norm": 0.6616294980049133,
      "learning_rate": 0.0004332658118737205,
      "loss": 3.5231,
      "step": 25900
    },
    {
      "epoch": 2.793025508556668,
      "grad_norm": 0.566481351852417,
      "learning_rate": 0.0004329425708436591,
      "loss": 3.5384,
      "step": 25950
    },
    {
      "epoch": 2.798407060596276,
      "grad_norm": 0.5798117518424988,
      "learning_rate": 0.00043261932981359763,
      "loss": 3.5461,
      "step": 26000
    },
    {
      "epoch": 2.798407060596276,
      "eval_accuracy": 0.3685674366643303,
      "eval_loss": 3.5130412578582764,
      "eval_runtime": 184.8813,
      "eval_samples_per_second": 97.419,
      "eval_steps_per_second": 6.09,
      "step": 26000
    },
    {
      "epoch": 2.803788612635884,
      "grad_norm": 0.5411988496780396,
      "learning_rate": 0.0004322960887835362,
      "loss": 3.5413,
      "step": 26050
    },
    {
      "epoch": 2.8091701646754923,
      "grad_norm": 0.554651141166687,
      "learning_rate": 0.00043197284775347476,
      "loss": 3.5336,
      "step": 26100
    },
    {
      "epoch": 2.8145517167151004,
      "grad_norm": 0.5864933133125305,
      "learning_rate": 0.00043164960672341336,
      "loss": 3.5489,
      "step": 26150
    },
    {
      "epoch": 2.819933268754709,
      "grad_norm": 0.549628734588623,
      "learning_rate": 0.000431326365693352,
      "loss": 3.534,
      "step": 26200
    },
    {
      "epoch": 2.825314820794317,
      "grad_norm": 0.5752499103546143,
      "learning_rate": 0.00043100312466329055,
      "loss": 3.5373,
      "step": 26250
    },
    {
      "epoch": 2.830696372833925,
      "grad_norm": 0.6056687831878662,
      "learning_rate": 0.00043067988363322914,
      "loss": 3.5467,
      "step": 26300
    },
    {
      "epoch": 2.8360779248735337,
      "grad_norm": 0.5557623505592346,
      "learning_rate": 0.00043035664260316774,
      "loss": 3.5347,
      "step": 26350
    },
    {
      "epoch": 2.841459476913142,
      "grad_norm": 0.5566017031669617,
      "learning_rate": 0.00043003986639370754,
      "loss": 3.5568,
      "step": 26400
    },
    {
      "epoch": 2.84684102895275,
      "grad_norm": 0.6013063192367554,
      "learning_rate": 0.0004297166253636461,
      "loss": 3.5329,
      "step": 26450
    },
    {
      "epoch": 2.852222580992358,
      "grad_norm": 0.560697078704834,
      "learning_rate": 0.0004293933843335847,
      "loss": 3.5274,
      "step": 26500
    },
    {
      "epoch": 2.857604133031966,
      "grad_norm": 0.5921282768249512,
      "learning_rate": 0.00042907014330352333,
      "loss": 3.538,
      "step": 26550
    },
    {
      "epoch": 2.8629856850715747,
      "grad_norm": 0.5770512223243713,
      "learning_rate": 0.00042874690227346187,
      "loss": 3.5341,
      "step": 26600
    },
    {
      "epoch": 2.868367237111183,
      "grad_norm": 0.5724422931671143,
      "learning_rate": 0.00042842366124340046,
      "loss": 3.5228,
      "step": 26650
    },
    {
      "epoch": 2.873748789150791,
      "grad_norm": 0.593283474445343,
      "learning_rate": 0.00042810042021333906,
      "loss": 3.535,
      "step": 26700
    },
    {
      "epoch": 2.8791303411903995,
      "grad_norm": 0.5896148681640625,
      "learning_rate": 0.0004277771791832776,
      "loss": 3.5122,
      "step": 26750
    },
    {
      "epoch": 2.8845118932300076,
      "grad_norm": 0.5596243143081665,
      "learning_rate": 0.0004274539381532162,
      "loss": 3.5354,
      "step": 26800
    },
    {
      "epoch": 2.8898934452696157,
      "grad_norm": 0.592008650302887,
      "learning_rate": 0.00042713069712315484,
      "loss": 3.5354,
      "step": 26850
    },
    {
      "epoch": 2.895274997309224,
      "grad_norm": 0.5777035355567932,
      "learning_rate": 0.0004268074560930934,
      "loss": 3.5294,
      "step": 26900
    },
    {
      "epoch": 2.9006565493488323,
      "grad_norm": 0.6243908405303955,
      "learning_rate": 0.000426484215063032,
      "loss": 3.5134,
      "step": 26950
    },
    {
      "epoch": 2.9060381013884404,
      "grad_norm": 0.5806100368499756,
      "learning_rate": 0.0004261609740329705,
      "loss": 3.5291,
      "step": 27000
    },
    {
      "epoch": 2.9060381013884404,
      "eval_accuracy": 0.36983183031982736,
      "eval_loss": 3.504441738128662,
      "eval_runtime": 185.054,
      "eval_samples_per_second": 97.328,
      "eval_steps_per_second": 6.085,
      "step": 27000
    },
    {
      "epoch": 2.9114196534280485,
      "grad_norm": 0.6097036600112915,
      "learning_rate": 0.0004258377330029091,
      "loss": 3.5165,
      "step": 27050
    },
    {
      "epoch": 2.9168012054676566,
      "grad_norm": 0.6032792925834656,
      "learning_rate": 0.00042551449197284776,
      "loss": 3.543,
      "step": 27100
    },
    {
      "epoch": 2.922182757507265,
      "grad_norm": 0.5959144830703735,
      "learning_rate": 0.0004251912509427863,
      "loss": 3.5216,
      "step": 27150
    },
    {
      "epoch": 2.9275643095468733,
      "grad_norm": 0.6524012684822083,
      "learning_rate": 0.0004248680099127249,
      "loss": 3.5105,
      "step": 27200
    },
    {
      "epoch": 2.9329458615864814,
      "grad_norm": 0.5885307788848877,
      "learning_rate": 0.0004245447688826635,
      "loss": 3.5274,
      "step": 27250
    },
    {
      "epoch": 2.93832741362609,
      "grad_norm": 0.6017599105834961,
      "learning_rate": 0.00042422152785260203,
      "loss": 3.5332,
      "step": 27300
    },
    {
      "epoch": 2.943708965665698,
      "grad_norm": 0.6001046895980835,
      "learning_rate": 0.0004238982868225406,
      "loss": 3.5223,
      "step": 27350
    },
    {
      "epoch": 2.949090517705306,
      "grad_norm": 0.5998550057411194,
      "learning_rate": 0.0004235750457924793,
      "loss": 3.5066,
      "step": 27400
    },
    {
      "epoch": 2.9544720697449143,
      "grad_norm": 0.6128471493721008,
      "learning_rate": 0.0004232518047624178,
      "loss": 3.5448,
      "step": 27450
    },
    {
      "epoch": 2.9598536217845224,
      "grad_norm": 0.5720791816711426,
      "learning_rate": 0.0004229285637323564,
      "loss": 3.5259,
      "step": 27500
    },
    {
      "epoch": 2.965235173824131,
      "grad_norm": 0.5852633118629456,
      "learning_rate": 0.00042260532270229495,
      "loss": 3.5241,
      "step": 27550
    },
    {
      "epoch": 2.970616725863739,
      "grad_norm": 0.6088577508926392,
      "learning_rate": 0.00042228208167223354,
      "loss": 3.518,
      "step": 27600
    },
    {
      "epoch": 2.975998277903347,
      "grad_norm": 0.5983795523643494,
      "learning_rate": 0.00042195884064217214,
      "loss": 3.5264,
      "step": 27650
    },
    {
      "epoch": 2.9813798299429557,
      "grad_norm": 0.5886030197143555,
      "learning_rate": 0.00042163559961211073,
      "loss": 3.5261,
      "step": 27700
    },
    {
      "epoch": 2.986761381982564,
      "grad_norm": 0.612271249294281,
      "learning_rate": 0.00042131235858204933,
      "loss": 3.5311,
      "step": 27750
    },
    {
      "epoch": 2.992142934022172,
      "grad_norm": 0.6254613399505615,
      "learning_rate": 0.0004209891175519879,
      "loss": 3.5097,
      "step": 27800
    },
    {
      "epoch": 2.9975244860617805,
      "grad_norm": 0.6479254364967346,
      "learning_rate": 0.00042066587652192646,
      "loss": 3.5238,
      "step": 27850
    },
    {
      "epoch": 3.0029060381013886,
      "grad_norm": 0.6095746755599976,
      "learning_rate": 0.00042034263549186506,
      "loss": 3.4804,
      "step": 27900
    },
    {
      "epoch": 3.0082875901409967,
      "grad_norm": 0.6377231478691101,
      "learning_rate": 0.0004200193944618036,
      "loss": 3.4438,
      "step": 27950
    },
    {
      "epoch": 3.0136691421806048,
      "grad_norm": 0.5965224504470825,
      "learning_rate": 0.00041969615343174225,
      "loss": 3.4265,
      "step": 28000
    },
    {
      "epoch": 3.0136691421806048,
      "eval_accuracy": 0.3701457285127234,
      "eval_loss": 3.5000498294830322,
      "eval_runtime": 184.7752,
      "eval_samples_per_second": 97.475,
      "eval_steps_per_second": 6.094,
      "step": 28000
    },
    {
      "epoch": 3.0190506942202133,
      "grad_norm": 0.5932812690734863,
      "learning_rate": 0.00041937291240168084,
      "loss": 3.4383,
      "step": 28050
    },
    {
      "epoch": 3.0244322462598214,
      "grad_norm": 0.6348116993904114,
      "learning_rate": 0.0004190496713716194,
      "loss": 3.4435,
      "step": 28100
    },
    {
      "epoch": 3.0298137982994295,
      "grad_norm": 0.623078465461731,
      "learning_rate": 0.000418726430341558,
      "loss": 3.4472,
      "step": 28150
    },
    {
      "epoch": 3.0351953503390376,
      "grad_norm": 0.5842487215995789,
      "learning_rate": 0.00041840318931149657,
      "loss": 3.4284,
      "step": 28200
    },
    {
      "epoch": 3.040576902378646,
      "grad_norm": 0.6037656664848328,
      "learning_rate": 0.00041807994828143517,
      "loss": 3.4273,
      "step": 28250
    },
    {
      "epoch": 3.0459584544182543,
      "grad_norm": 0.5624446868896484,
      "learning_rate": 0.00041775670725137376,
      "loss": 3.4341,
      "step": 28300
    },
    {
      "epoch": 3.0513400064578624,
      "grad_norm": 0.6236454248428345,
      "learning_rate": 0.00041743346622131236,
      "loss": 3.4384,
      "step": 28350
    },
    {
      "epoch": 3.0567215584974705,
      "grad_norm": 0.5729033350944519,
      "learning_rate": 0.0004171102251912509,
      "loss": 3.4509,
      "step": 28400
    },
    {
      "epoch": 3.062103110537079,
      "grad_norm": 0.570470929145813,
      "learning_rate": 0.0004167869841611895,
      "loss": 3.4454,
      "step": 28450
    },
    {
      "epoch": 3.067484662576687,
      "grad_norm": 0.5834895372390747,
      "learning_rate": 0.00041646374313112803,
      "loss": 3.4471,
      "step": 28500
    },
    {
      "epoch": 3.0728662146162953,
      "grad_norm": 0.6381838917732239,
      "learning_rate": 0.0004161405021010667,
      "loss": 3.4475,
      "step": 28550
    },
    {
      "epoch": 3.0782477666559034,
      "grad_norm": 0.611271321773529,
      "learning_rate": 0.00041582372589160643,
      "loss": 3.452,
      "step": 28600
    },
    {
      "epoch": 3.083629318695512,
      "grad_norm": 0.6530306339263916,
      "learning_rate": 0.0004155004848615451,
      "loss": 3.4317,
      "step": 28650
    },
    {
      "epoch": 3.08901087073512,
      "grad_norm": 0.6146801710128784,
      "learning_rate": 0.0004151772438314837,
      "loss": 3.4379,
      "step": 28700
    },
    {
      "epoch": 3.094392422774728,
      "grad_norm": 0.5884636640548706,
      "learning_rate": 0.0004148540028014222,
      "loss": 3.4381,
      "step": 28750
    },
    {
      "epoch": 3.0997739748143363,
      "grad_norm": 0.6561384797096252,
      "learning_rate": 0.0004145307617713608,
      "loss": 3.4375,
      "step": 28800
    },
    {
      "epoch": 3.105155526853945,
      "grad_norm": 0.5567984580993652,
      "learning_rate": 0.00041420752074129935,
      "loss": 3.4381,
      "step": 28850
    },
    {
      "epoch": 3.110537078893553,
      "grad_norm": 0.583527684211731,
      "learning_rate": 0.000413884279711238,
      "loss": 3.4351,
      "step": 28900
    },
    {
      "epoch": 3.115918630933161,
      "grad_norm": 0.580026388168335,
      "learning_rate": 0.0004135610386811766,
      "loss": 3.443,
      "step": 28950
    },
    {
      "epoch": 3.121300182972769,
      "grad_norm": 0.6263661980628967,
      "learning_rate": 0.00041323779765111514,
      "loss": 3.4472,
      "step": 29000
    },
    {
      "epoch": 3.121300182972769,
      "eval_accuracy": 0.37119868364853353,
      "eval_loss": 3.4976308345794678,
      "eval_runtime": 184.8903,
      "eval_samples_per_second": 97.415,
      "eval_steps_per_second": 6.09,
      "step": 29000
    },
    {
      "epoch": 3.1266817350123777,
      "grad_norm": 0.5766123533248901,
      "learning_rate": 0.00041291455662105373,
      "loss": 3.4438,
      "step": 29050
    },
    {
      "epoch": 3.132063287051986,
      "grad_norm": 0.5877207517623901,
      "learning_rate": 0.0004125913155909923,
      "loss": 3.4321,
      "step": 29100
    },
    {
      "epoch": 3.137444839091594,
      "grad_norm": 0.6023469567298889,
      "learning_rate": 0.00041226807456093087,
      "loss": 3.4396,
      "step": 29150
    },
    {
      "epoch": 3.1428263911312024,
      "grad_norm": 0.5752248167991638,
      "learning_rate": 0.0004119448335308695,
      "loss": 3.4291,
      "step": 29200
    },
    {
      "epoch": 3.1482079431708105,
      "grad_norm": 0.5830261707305908,
      "learning_rate": 0.0004116215925008081,
      "loss": 3.4434,
      "step": 29250
    },
    {
      "epoch": 3.1535894952104186,
      "grad_norm": 0.6252485513687134,
      "learning_rate": 0.00041129835147074665,
      "loss": 3.458,
      "step": 29300
    },
    {
      "epoch": 3.1589710472500268,
      "grad_norm": 0.6096091866493225,
      "learning_rate": 0.00041097511044068524,
      "loss": 3.4718,
      "step": 29350
    },
    {
      "epoch": 3.1643525992896353,
      "grad_norm": 0.6204565167427063,
      "learning_rate": 0.0004106518694106238,
      "loss": 3.4728,
      "step": 29400
    },
    {
      "epoch": 3.1697341513292434,
      "grad_norm": 0.6110973954200745,
      "learning_rate": 0.0004103286283805624,
      "loss": 3.4484,
      "step": 29450
    },
    {
      "epoch": 3.1751157033688515,
      "grad_norm": 0.6513490080833435,
      "learning_rate": 0.00041000538735050103,
      "loss": 3.4548,
      "step": 29500
    },
    {
      "epoch": 3.1804972554084596,
      "grad_norm": 0.6093899607658386,
      "learning_rate": 0.00040968214632043957,
      "loss": 3.4613,
      "step": 29550
    },
    {
      "epoch": 3.185878807448068,
      "grad_norm": 0.6108792424201965,
      "learning_rate": 0.00040935890529037816,
      "loss": 3.4307,
      "step": 29600
    },
    {
      "epoch": 3.1912603594876763,
      "grad_norm": 0.6220118999481201,
      "learning_rate": 0.00040903566426031676,
      "loss": 3.4599,
      "step": 29650
    },
    {
      "epoch": 3.1966419115272844,
      "grad_norm": 0.5830521583557129,
      "learning_rate": 0.0004087124232302553,
      "loss": 3.4543,
      "step": 29700
    },
    {
      "epoch": 3.2020234635668925,
      "grad_norm": 0.6141335368156433,
      "learning_rate": 0.0004083891822001939,
      "loss": 3.4441,
      "step": 29750
    },
    {
      "epoch": 3.207405015606501,
      "grad_norm": 0.6098508238792419,
      "learning_rate": 0.00040806594117013254,
      "loss": 3.4538,
      "step": 29800
    },
    {
      "epoch": 3.212786567646109,
      "grad_norm": 0.5632219314575195,
      "learning_rate": 0.0004077427001400711,
      "loss": 3.4533,
      "step": 29850
    },
    {
      "epoch": 3.2181681196857173,
      "grad_norm": 0.6396539807319641,
      "learning_rate": 0.0004074194591100097,
      "loss": 3.4618,
      "step": 29900
    },
    {
      "epoch": 3.2235496717253254,
      "grad_norm": 0.6783267855644226,
      "learning_rate": 0.0004070962180799482,
      "loss": 3.4458,
      "step": 29950
    },
    {
      "epoch": 3.228931223764934,
      "grad_norm": 0.6168944239616394,
      "learning_rate": 0.0004067729770498868,
      "loss": 3.4667,
      "step": 30000
    },
    {
      "epoch": 3.228931223764934,
      "eval_accuracy": 0.37168392744551304,
      "eval_loss": 3.4887137413024902,
      "eval_runtime": 184.9069,
      "eval_samples_per_second": 97.406,
      "eval_steps_per_second": 6.09,
      "step": 30000
    },
    {
      "epoch": 3.234312775804542,
      "grad_norm": 0.6674462556838989,
      "learning_rate": 0.00040644973601982546,
      "loss": 3.4648,
      "step": 30050
    },
    {
      "epoch": 3.23969432784415,
      "grad_norm": 0.5786466598510742,
      "learning_rate": 0.000406126494989764,
      "loss": 3.4439,
      "step": 30100
    },
    {
      "epoch": 3.2450758798837587,
      "grad_norm": 0.6295211315155029,
      "learning_rate": 0.0004058032539597026,
      "loss": 3.4525,
      "step": 30150
    },
    {
      "epoch": 3.250457431923367,
      "grad_norm": 0.659146249294281,
      "learning_rate": 0.00040548001292964114,
      "loss": 3.4615,
      "step": 30200
    },
    {
      "epoch": 3.255838983962975,
      "grad_norm": 0.5946276187896729,
      "learning_rate": 0.00040515677189957973,
      "loss": 3.4612,
      "step": 30250
    },
    {
      "epoch": 3.261220536002583,
      "grad_norm": 0.6724011898040771,
      "learning_rate": 0.0004048335308695183,
      "loss": 3.4394,
      "step": 30300
    },
    {
      "epoch": 3.2666020880421915,
      "grad_norm": 0.6301489472389221,
      "learning_rate": 0.0004045102898394569,
      "loss": 3.4501,
      "step": 30350
    },
    {
      "epoch": 3.2719836400817996,
      "grad_norm": 0.666102945804596,
      "learning_rate": 0.0004041870488093955,
      "loss": 3.4573,
      "step": 30400
    },
    {
      "epoch": 3.2773651921214078,
      "grad_norm": 0.6429094672203064,
      "learning_rate": 0.0004038638077793341,
      "loss": 3.4675,
      "step": 30450
    },
    {
      "epoch": 3.282746744161016,
      "grad_norm": 0.6529164910316467,
      "learning_rate": 0.00040354056674927265,
      "loss": 3.4523,
      "step": 30500
    },
    {
      "epoch": 3.2881282962006244,
      "grad_norm": 0.593303918838501,
      "learning_rate": 0.00040321732571921124,
      "loss": 3.4455,
      "step": 30550
    },
    {
      "epoch": 3.2935098482402325,
      "grad_norm": 0.587743878364563,
      "learning_rate": 0.0004029070143303523,
      "loss": 3.4432,
      "step": 30600
    },
    {
      "epoch": 3.2988914002798406,
      "grad_norm": 0.6212865114212036,
      "learning_rate": 0.00040258377330029086,
      "loss": 3.4705,
      "step": 30650
    },
    {
      "epoch": 3.304272952319449,
      "grad_norm": 0.5814943313598633,
      "learning_rate": 0.00040226053227022945,
      "loss": 3.4567,
      "step": 30700
    },
    {
      "epoch": 3.3096545043590573,
      "grad_norm": 0.5858567953109741,
      "learning_rate": 0.00040193729124016805,
      "loss": 3.453,
      "step": 30750
    },
    {
      "epoch": 3.3150360563986654,
      "grad_norm": 0.6212705373764038,
      "learning_rate": 0.00040161405021010664,
      "loss": 3.4501,
      "step": 30800
    },
    {
      "epoch": 3.3204176084382735,
      "grad_norm": 0.6279528141021729,
      "learning_rate": 0.00040129080918004524,
      "loss": 3.4556,
      "step": 30850
    },
    {
      "epoch": 3.3257991604778816,
      "grad_norm": 0.6483830213546753,
      "learning_rate": 0.00040096756814998383,
      "loss": 3.4473,
      "step": 30900
    },
    {
      "epoch": 3.33118071251749,
      "grad_norm": 0.6192697286605835,
      "learning_rate": 0.00040064432711992237,
      "loss": 3.47,
      "step": 30950
    },
    {
      "epoch": 3.3365622645570983,
      "grad_norm": 0.6230787038803101,
      "learning_rate": 0.00040032108608986097,
      "loss": 3.4857,
      "step": 31000
    },
    {
      "epoch": 3.3365622645570983,
      "eval_accuracy": 0.3728218491395941,
      "eval_loss": 3.4819231033325195,
      "eval_runtime": 185.0123,
      "eval_samples_per_second": 97.35,
      "eval_steps_per_second": 6.086,
      "step": 31000
    },
    {
      "epoch": 3.3419438165967064,
      "grad_norm": 0.6070190072059631,
      "learning_rate": 0.0003999978450597995,
      "loss": 3.4448,
      "step": 31050
    },
    {
      "epoch": 3.347325368636315,
      "grad_norm": 0.638231098651886,
      "learning_rate": 0.00039967460402973816,
      "loss": 3.4718,
      "step": 31100
    },
    {
      "epoch": 3.352706920675923,
      "grad_norm": 0.7029089331626892,
      "learning_rate": 0.00039935136299967675,
      "loss": 3.4692,
      "step": 31150
    },
    {
      "epoch": 3.358088472715531,
      "grad_norm": 0.565448522567749,
      "learning_rate": 0.0003990281219696153,
      "loss": 3.4601,
      "step": 31200
    },
    {
      "epoch": 3.3634700247551392,
      "grad_norm": 0.6335359811782837,
      "learning_rate": 0.0003987048809395539,
      "loss": 3.4573,
      "step": 31250
    },
    {
      "epoch": 3.368851576794748,
      "grad_norm": 0.604525625705719,
      "learning_rate": 0.0003983816399094925,
      "loss": 3.4508,
      "step": 31300
    },
    {
      "epoch": 3.374233128834356,
      "grad_norm": 0.634780764579773,
      "learning_rate": 0.000398058398879431,
      "loss": 3.4762,
      "step": 31350
    },
    {
      "epoch": 3.379614680873964,
      "grad_norm": 0.636223316192627,
      "learning_rate": 0.00039773515784936967,
      "loss": 3.4521,
      "step": 31400
    },
    {
      "epoch": 3.384996232913572,
      "grad_norm": 0.6109070181846619,
      "learning_rate": 0.00039741191681930826,
      "loss": 3.4659,
      "step": 31450
    },
    {
      "epoch": 3.3903777849531807,
      "grad_norm": 0.6315115094184875,
      "learning_rate": 0.0003970886757892468,
      "loss": 3.4598,
      "step": 31500
    },
    {
      "epoch": 3.3957593369927888,
      "grad_norm": 0.5966545343399048,
      "learning_rate": 0.0003967654347591854,
      "loss": 3.458,
      "step": 31550
    },
    {
      "epoch": 3.401140889032397,
      "grad_norm": 0.6154278516769409,
      "learning_rate": 0.00039644219372912394,
      "loss": 3.4428,
      "step": 31600
    },
    {
      "epoch": 3.4065224410720054,
      "grad_norm": 0.6580645442008972,
      "learning_rate": 0.0003961189526990626,
      "loss": 3.4463,
      "step": 31650
    },
    {
      "epoch": 3.4119039931116135,
      "grad_norm": 0.6818550825119019,
      "learning_rate": 0.0003957957116690012,
      "loss": 3.4425,
      "step": 31700
    },
    {
      "epoch": 3.4172855451512216,
      "grad_norm": 0.6167941689491272,
      "learning_rate": 0.0003954724706389397,
      "loss": 3.4583,
      "step": 31750
    },
    {
      "epoch": 3.4226670971908297,
      "grad_norm": 0.6051561236381531,
      "learning_rate": 0.0003951492296088783,
      "loss": 3.4626,
      "step": 31800
    },
    {
      "epoch": 3.428048649230438,
      "grad_norm": 0.604656457901001,
      "learning_rate": 0.0003948259885788169,
      "loss": 3.456,
      "step": 31850
    },
    {
      "epoch": 3.4334302012700464,
      "grad_norm": 0.6075143814086914,
      "learning_rate": 0.00039450274754875545,
      "loss": 3.4499,
      "step": 31900
    },
    {
      "epoch": 3.4388117533096545,
      "grad_norm": 0.6349363327026367,
      "learning_rate": 0.0003941795065186941,
      "loss": 3.4608,
      "step": 31950
    },
    {
      "epoch": 3.4441933053492626,
      "grad_norm": 0.6251042485237122,
      "learning_rate": 0.0003938562654886327,
      "loss": 3.4601,
      "step": 32000
    },
    {
      "epoch": 3.4441933053492626,
      "eval_accuracy": 0.372848251791333,
      "eval_loss": 3.476959466934204,
      "eval_runtime": 185.0651,
      "eval_samples_per_second": 97.323,
      "eval_steps_per_second": 6.084,
      "step": 32000
    },
    {
      "epoch": 3.449574857388871,
      "grad_norm": 0.6375144720077515,
      "learning_rate": 0.00039353302445857124,
      "loss": 3.4643,
      "step": 32050
    },
    {
      "epoch": 3.4549564094284793,
      "grad_norm": 0.6365067958831787,
      "learning_rate": 0.00039320978342850983,
      "loss": 3.4534,
      "step": 32100
    },
    {
      "epoch": 3.4603379614680874,
      "grad_norm": 0.6374215483665466,
      "learning_rate": 0.00039288654239844837,
      "loss": 3.4464,
      "step": 32150
    },
    {
      "epoch": 3.4657195135076955,
      "grad_norm": 0.6080241203308105,
      "learning_rate": 0.00039256330136838697,
      "loss": 3.4582,
      "step": 32200
    },
    {
      "epoch": 3.471101065547304,
      "grad_norm": 0.5984114408493042,
      "learning_rate": 0.0003922400603383256,
      "loss": 3.456,
      "step": 32250
    },
    {
      "epoch": 3.476482617586912,
      "grad_norm": 0.6118507981300354,
      "learning_rate": 0.00039191681930826416,
      "loss": 3.4786,
      "step": 32300
    },
    {
      "epoch": 3.4818641696265202,
      "grad_norm": 0.6618357300758362,
      "learning_rate": 0.00039159357827820275,
      "loss": 3.4559,
      "step": 32350
    },
    {
      "epoch": 3.4872457216661283,
      "grad_norm": 0.6183602809906006,
      "learning_rate": 0.00039127033724814135,
      "loss": 3.462,
      "step": 32400
    },
    {
      "epoch": 3.492627273705737,
      "grad_norm": 0.6030476689338684,
      "learning_rate": 0.0003909470962180799,
      "loss": 3.4661,
      "step": 32450
    },
    {
      "epoch": 3.498008825745345,
      "grad_norm": 0.6316002011299133,
      "learning_rate": 0.00039062385518801854,
      "loss": 3.4361,
      "step": 32500
    },
    {
      "epoch": 3.503390377784953,
      "grad_norm": 0.592577338218689,
      "learning_rate": 0.00039030061415795713,
      "loss": 3.438,
      "step": 32550
    },
    {
      "epoch": 3.5087719298245617,
      "grad_norm": 0.6018152236938477,
      "learning_rate": 0.00038997737312789567,
      "loss": 3.4465,
      "step": 32600
    },
    {
      "epoch": 3.5141534818641698,
      "grad_norm": 0.657094419002533,
      "learning_rate": 0.00038965413209783426,
      "loss": 3.4449,
      "step": 32650
    },
    {
      "epoch": 3.519535033903778,
      "grad_norm": 0.6502700448036194,
      "learning_rate": 0.0003893308910677728,
      "loss": 3.4711,
      "step": 32700
    },
    {
      "epoch": 3.524916585943386,
      "grad_norm": 0.6325830817222595,
      "learning_rate": 0.0003890076500377114,
      "loss": 3.4545,
      "step": 32750
    },
    {
      "epoch": 3.530298137982994,
      "grad_norm": 0.6116040349006653,
      "learning_rate": 0.00038868440900765005,
      "loss": 3.4498,
      "step": 32800
    },
    {
      "epoch": 3.5356796900226026,
      "grad_norm": 0.6353474855422974,
      "learning_rate": 0.0003883611679775886,
      "loss": 3.4359,
      "step": 32850
    },
    {
      "epoch": 3.5410612420622107,
      "grad_norm": 0.6394508481025696,
      "learning_rate": 0.0003880379269475272,
      "loss": 3.4538,
      "step": 32900
    },
    {
      "epoch": 3.546442794101819,
      "grad_norm": 0.5920386910438538,
      "learning_rate": 0.0003877146859174657,
      "loss": 3.4638,
      "step": 32950
    },
    {
      "epoch": 3.5518243461414274,
      "grad_norm": 0.6352172493934631,
      "learning_rate": 0.0003873914448874043,
      "loss": 3.4542,
      "step": 33000
    },
    {
      "epoch": 3.5518243461414274,
      "eval_accuracy": 0.3739409738840832,
      "eval_loss": 3.467642068862915,
      "eval_runtime": 184.7948,
      "eval_samples_per_second": 97.465,
      "eval_steps_per_second": 6.093,
      "step": 33000
    },
    {
      "epoch": 3.5572058981810355,
      "grad_norm": 0.5835293531417847,
      "learning_rate": 0.0003870682038573429,
      "loss": 3.4507,
      "step": 33050
    },
    {
      "epoch": 3.5625874502206436,
      "grad_norm": 0.6255052089691162,
      "learning_rate": 0.0003867449628272815,
      "loss": 3.4522,
      "step": 33100
    },
    {
      "epoch": 3.5679690022602517,
      "grad_norm": 0.6663792133331299,
      "learning_rate": 0.0003864217217972201,
      "loss": 3.4373,
      "step": 33150
    },
    {
      "epoch": 3.57335055429986,
      "grad_norm": 0.5987926125526428,
      "learning_rate": 0.0003860984807671587,
      "loss": 3.4538,
      "step": 33200
    },
    {
      "epoch": 3.5787321063394684,
      "grad_norm": 0.6144534945487976,
      "learning_rate": 0.00038577523973709724,
      "loss": 3.466,
      "step": 33250
    },
    {
      "epoch": 3.5841136583790765,
      "grad_norm": 0.6252511739730835,
      "learning_rate": 0.00038545199870703583,
      "loss": 3.4454,
      "step": 33300
    },
    {
      "epoch": 3.5894952104186846,
      "grad_norm": 0.6555866599082947,
      "learning_rate": 0.0003851287576769744,
      "loss": 3.455,
      "step": 33350
    },
    {
      "epoch": 3.594876762458293,
      "grad_norm": 0.6036868691444397,
      "learning_rate": 0.000384805516646913,
      "loss": 3.4569,
      "step": 33400
    },
    {
      "epoch": 3.6002583144979012,
      "grad_norm": 0.6310072541236877,
      "learning_rate": 0.0003844822756168516,
      "loss": 3.4375,
      "step": 33450
    },
    {
      "epoch": 3.6056398665375093,
      "grad_norm": 0.6304979920387268,
      "learning_rate": 0.00038415903458679016,
      "loss": 3.4567,
      "step": 33500
    },
    {
      "epoch": 3.611021418577118,
      "grad_norm": 0.6037408113479614,
      "learning_rate": 0.00038383579355672875,
      "loss": 3.4643,
      "step": 33550
    },
    {
      "epoch": 3.616402970616726,
      "grad_norm": 0.6693500280380249,
      "learning_rate": 0.00038351255252666735,
      "loss": 3.4298,
      "step": 33600
    },
    {
      "epoch": 3.621784522656334,
      "grad_norm": 0.6045053005218506,
      "learning_rate": 0.00038318931149660594,
      "loss": 3.4583,
      "step": 33650
    },
    {
      "epoch": 3.627166074695942,
      "grad_norm": 0.5890701413154602,
      "learning_rate": 0.00038286607046654454,
      "loss": 3.4497,
      "step": 33700
    },
    {
      "epoch": 3.6325476267355503,
      "grad_norm": 0.6815541386604309,
      "learning_rate": 0.00038254282943648313,
      "loss": 3.4522,
      "step": 33750
    },
    {
      "epoch": 3.637929178775159,
      "grad_norm": 0.6552605032920837,
      "learning_rate": 0.00038221958840642167,
      "loss": 3.466,
      "step": 33800
    },
    {
      "epoch": 3.643310730814767,
      "grad_norm": 0.6101818084716797,
      "learning_rate": 0.00038189634737636027,
      "loss": 3.4601,
      "step": 33850
    },
    {
      "epoch": 3.648692282854375,
      "grad_norm": 0.636333703994751,
      "learning_rate": 0.0003815731063462988,
      "loss": 3.4451,
      "step": 33900
    },
    {
      "epoch": 3.6540738348939836,
      "grad_norm": 0.6014490723609924,
      "learning_rate": 0.00038124986531623745,
      "loss": 3.4576,
      "step": 33950
    },
    {
      "epoch": 3.6594553869335917,
      "grad_norm": 0.6211981773376465,
      "learning_rate": 0.00038092662428617605,
      "loss": 3.4527,
      "step": 34000
    },
    {
      "epoch": 3.6594553869335917,
      "eval_accuracy": 0.374348748172051,
      "eval_loss": 3.4629900455474854,
      "eval_runtime": 184.9732,
      "eval_samples_per_second": 97.371,
      "eval_steps_per_second": 6.087,
      "step": 34000
    },
    {
      "epoch": 3.6648369389732,
      "grad_norm": 0.65215665102005,
      "learning_rate": 0.0003806033832561146,
      "loss": 3.4392,
      "step": 34050
    },
    {
      "epoch": 3.670218491012808,
      "grad_norm": 0.6055262684822083,
      "learning_rate": 0.0003802801422260532,
      "loss": 3.4235,
      "step": 34100
    },
    {
      "epoch": 3.675600043052416,
      "grad_norm": 0.6137178540229797,
      "learning_rate": 0.0003799569011959918,
      "loss": 3.4482,
      "step": 34150
    },
    {
      "epoch": 3.6809815950920246,
      "grad_norm": 0.6887802481651306,
      "learning_rate": 0.0003796336601659303,
      "loss": 3.4466,
      "step": 34200
    },
    {
      "epoch": 3.6863631471316327,
      "grad_norm": 0.634081244468689,
      "learning_rate": 0.00037931041913586897,
      "loss": 3.4466,
      "step": 34250
    },
    {
      "epoch": 3.691744699171241,
      "grad_norm": 0.622760534286499,
      "learning_rate": 0.00037898717810580756,
      "loss": 3.4431,
      "step": 34300
    },
    {
      "epoch": 3.6971262512108494,
      "grad_norm": 0.616561770439148,
      "learning_rate": 0.0003786639370757461,
      "loss": 3.454,
      "step": 34350
    },
    {
      "epoch": 3.7025078032504575,
      "grad_norm": 0.6342828273773193,
      "learning_rate": 0.0003783406960456847,
      "loss": 3.4736,
      "step": 34400
    },
    {
      "epoch": 3.7078893552900656,
      "grad_norm": 0.7093660235404968,
      "learning_rate": 0.00037801745501562324,
      "loss": 3.4731,
      "step": 34450
    },
    {
      "epoch": 3.713270907329674,
      "grad_norm": 0.6421014666557312,
      "learning_rate": 0.0003776942139855619,
      "loss": 3.4593,
      "step": 34500
    },
    {
      "epoch": 3.7186524593692822,
      "grad_norm": 0.6160980463027954,
      "learning_rate": 0.0003773709729555005,
      "loss": 3.4535,
      "step": 34550
    },
    {
      "epoch": 3.7240340114088903,
      "grad_norm": 0.6290106177330017,
      "learning_rate": 0.000377047731925439,
      "loss": 3.461,
      "step": 34600
    },
    {
      "epoch": 3.7294155634484984,
      "grad_norm": 0.6211118698120117,
      "learning_rate": 0.0003767244908953776,
      "loss": 3.4576,
      "step": 34650
    },
    {
      "epoch": 3.7347971154881066,
      "grad_norm": 0.6442049145698547,
      "learning_rate": 0.0003764077146859174,
      "loss": 3.4573,
      "step": 34700
    },
    {
      "epoch": 3.740178667527715,
      "grad_norm": 0.6050986647605896,
      "learning_rate": 0.000376084473655856,
      "loss": 3.4425,
      "step": 34750
    },
    {
      "epoch": 3.745560219567323,
      "grad_norm": 0.6353182792663574,
      "learning_rate": 0.00037576123262579456,
      "loss": 3.437,
      "step": 34800
    },
    {
      "epoch": 3.7509417716069313,
      "grad_norm": 0.6667701005935669,
      "learning_rate": 0.00037543799159573315,
      "loss": 3.4571,
      "step": 34850
    },
    {
      "epoch": 3.75632332364654,
      "grad_norm": 0.6093037128448486,
      "learning_rate": 0.0003751147505656718,
      "loss": 3.4453,
      "step": 34900
    },
    {
      "epoch": 3.761704875686148,
      "grad_norm": 0.5967413783073425,
      "learning_rate": 0.00037479150953561034,
      "loss": 3.4659,
      "step": 34950
    },
    {
      "epoch": 3.767086427725756,
      "grad_norm": 0.6182402968406677,
      "learning_rate": 0.00037446826850554894,
      "loss": 3.4603,
      "step": 35000
    },
    {
      "epoch": 3.767086427725756,
      "eval_accuracy": 0.37532618955083,
      "eval_loss": 3.454987049102783,
      "eval_runtime": 184.7797,
      "eval_samples_per_second": 97.473,
      "eval_steps_per_second": 6.094,
      "step": 35000
    },
    {
      "epoch": 3.772467979765364,
      "grad_norm": 0.6796460747718811,
      "learning_rate": 0.00037414502747548753,
      "loss": 3.4427,
      "step": 35050
    },
    {
      "epoch": 3.7778495318049723,
      "grad_norm": 0.6759485006332397,
      "learning_rate": 0.00037382178644542607,
      "loss": 3.4735,
      "step": 35100
    },
    {
      "epoch": 3.783231083844581,
      "grad_norm": 0.6431259512901306,
      "learning_rate": 0.00037349854541536467,
      "loss": 3.4669,
      "step": 35150
    },
    {
      "epoch": 3.788612635884189,
      "grad_norm": 0.6208698153495789,
      "learning_rate": 0.0003731753043853033,
      "loss": 3.4517,
      "step": 35200
    },
    {
      "epoch": 3.793994187923797,
      "grad_norm": 0.6531832218170166,
      "learning_rate": 0.00037285206335524186,
      "loss": 3.459,
      "step": 35250
    },
    {
      "epoch": 3.7993757399634056,
      "grad_norm": 0.6968291997909546,
      "learning_rate": 0.00037252882232518045,
      "loss": 3.45,
      "step": 35300
    },
    {
      "epoch": 3.8047572920030137,
      "grad_norm": 0.6449758410453796,
      "learning_rate": 0.000372205581295119,
      "loss": 3.4481,
      "step": 35350
    },
    {
      "epoch": 3.810138844042622,
      "grad_norm": 0.6235898733139038,
      "learning_rate": 0.0003718823402650576,
      "loss": 3.4659,
      "step": 35400
    },
    {
      "epoch": 3.8155203960822304,
      "grad_norm": 0.7655190825462341,
      "learning_rate": 0.00037155909923499624,
      "loss": 3.4616,
      "step": 35450
    },
    {
      "epoch": 3.8209019481218385,
      "grad_norm": 0.6297181248664856,
      "learning_rate": 0.0003712358582049348,
      "loss": 3.4774,
      "step": 35500
    },
    {
      "epoch": 3.8262835001614466,
      "grad_norm": 0.6621859073638916,
      "learning_rate": 0.00037091261717487337,
      "loss": 3.4521,
      "step": 35550
    },
    {
      "epoch": 3.8316650522010547,
      "grad_norm": 0.5900010466575623,
      "learning_rate": 0.00037058937614481197,
      "loss": 3.447,
      "step": 35600
    },
    {
      "epoch": 3.837046604240663,
      "grad_norm": 0.6074267625808716,
      "learning_rate": 0.0003702661351147505,
      "loss": 3.4478,
      "step": 35650
    },
    {
      "epoch": 3.8424281562802713,
      "grad_norm": 0.616184413433075,
      "learning_rate": 0.0003699428940846891,
      "loss": 3.431,
      "step": 35700
    },
    {
      "epoch": 3.8478097083198795,
      "grad_norm": 0.6139175891876221,
      "learning_rate": 0.00036961965305462775,
      "loss": 3.4688,
      "step": 35750
    },
    {
      "epoch": 3.8531912603594876,
      "grad_norm": 0.6691179871559143,
      "learning_rate": 0.0003692964120245663,
      "loss": 3.4461,
      "step": 35800
    },
    {
      "epoch": 3.858572812399096,
      "grad_norm": 0.6166695356369019,
      "learning_rate": 0.0003689731709945049,
      "loss": 3.4642,
      "step": 35850
    },
    {
      "epoch": 3.863954364438704,
      "grad_norm": 0.6620895266532898,
      "learning_rate": 0.0003686499299644434,
      "loss": 3.4599,
      "step": 35900
    },
    {
      "epoch": 3.8693359164783123,
      "grad_norm": 0.58867347240448,
      "learning_rate": 0.000368326688934382,
      "loss": 3.4494,
      "step": 35950
    },
    {
      "epoch": 3.8747174685179204,
      "grad_norm": 0.6114178895950317,
      "learning_rate": 0.0003680034479043206,
      "loss": 3.4545,
      "step": 36000
    },
    {
      "epoch": 3.8747174685179204,
      "eval_accuracy": 0.37561944369504136,
      "eval_loss": 3.451740026473999,
      "eval_runtime": 184.9248,
      "eval_samples_per_second": 97.396,
      "eval_steps_per_second": 6.089,
      "step": 36000
    },
    {
      "epoch": 3.8800990205575285,
      "grad_norm": 0.6434826850891113,
      "learning_rate": 0.0003676802068742592,
      "loss": 3.4546,
      "step": 36050
    },
    {
      "epoch": 3.885480572597137,
      "grad_norm": 0.6056340336799622,
      "learning_rate": 0.0003673569658441978,
      "loss": 3.4587,
      "step": 36100
    },
    {
      "epoch": 3.890862124636745,
      "grad_norm": 0.6036874055862427,
      "learning_rate": 0.0003670337248141364,
      "loss": 3.4457,
      "step": 36150
    },
    {
      "epoch": 3.8962436766763533,
      "grad_norm": 0.6100910305976868,
      "learning_rate": 0.00036671048378407494,
      "loss": 3.4508,
      "step": 36200
    },
    {
      "epoch": 3.901625228715962,
      "grad_norm": 0.6239060759544373,
      "learning_rate": 0.00036638724275401353,
      "loss": 3.4465,
      "step": 36250
    },
    {
      "epoch": 3.90700678075557,
      "grad_norm": 0.6256414651870728,
      "learning_rate": 0.00036607046654455334,
      "loss": 3.4362,
      "step": 36300
    },
    {
      "epoch": 3.912388332795178,
      "grad_norm": 0.5970678329467773,
      "learning_rate": 0.00036574722551449193,
      "loss": 3.4634,
      "step": 36350
    },
    {
      "epoch": 3.9177698848347866,
      "grad_norm": 0.6506580114364624,
      "learning_rate": 0.00036542398448443053,
      "loss": 3.4448,
      "step": 36400
    },
    {
      "epoch": 3.9231514368743947,
      "grad_norm": 0.64057856798172,
      "learning_rate": 0.0003651007434543691,
      "loss": 3.445,
      "step": 36450
    },
    {
      "epoch": 3.928532988914003,
      "grad_norm": 0.6665002703666687,
      "learning_rate": 0.0003647775024243077,
      "loss": 3.4366,
      "step": 36500
    },
    {
      "epoch": 3.933914540953611,
      "grad_norm": 0.6165843605995178,
      "learning_rate": 0.00036445426139424626,
      "loss": 3.4316,
      "step": 36550
    },
    {
      "epoch": 3.939296092993219,
      "grad_norm": 0.6074864268302917,
      "learning_rate": 0.00036413102036418485,
      "loss": 3.4501,
      "step": 36600
    },
    {
      "epoch": 3.9446776450328276,
      "grad_norm": 0.7129690647125244,
      "learning_rate": 0.0003638077793341234,
      "loss": 3.4409,
      "step": 36650
    },
    {
      "epoch": 3.9500591970724357,
      "grad_norm": 0.6536222100257874,
      "learning_rate": 0.00036348453830406204,
      "loss": 3.4472,
      "step": 36700
    },
    {
      "epoch": 3.955440749112044,
      "grad_norm": 0.5931937098503113,
      "learning_rate": 0.00036316129727400064,
      "loss": 3.4439,
      "step": 36750
    },
    {
      "epoch": 3.9608223011516523,
      "grad_norm": 0.6528921127319336,
      "learning_rate": 0.0003628380562439392,
      "loss": 3.4427,
      "step": 36800
    },
    {
      "epoch": 3.9662038531912605,
      "grad_norm": 0.5929076075553894,
      "learning_rate": 0.00036251481521387777,
      "loss": 3.4536,
      "step": 36850
    },
    {
      "epoch": 3.9715854052308686,
      "grad_norm": 0.6636005640029907,
      "learning_rate": 0.00036219157418381637,
      "loss": 3.4588,
      "step": 36900
    },
    {
      "epoch": 3.9769669572704767,
      "grad_norm": 0.5989009737968445,
      "learning_rate": 0.0003618683331537549,
      "loss": 3.4483,
      "step": 36950
    },
    {
      "epoch": 3.9823485093100848,
      "grad_norm": 0.6094232201576233,
      "learning_rate": 0.00036154509212369356,
      "loss": 3.456,
      "step": 37000
    },
    {
      "epoch": 3.9823485093100848,
      "eval_accuracy": 0.37639761567756863,
      "eval_loss": 3.444646120071411,
      "eval_runtime": 184.7393,
      "eval_samples_per_second": 97.494,
      "eval_steps_per_second": 6.095,
      "step": 37000
    },
    {
      "epoch": 3.9877300613496933,
      "grad_norm": 0.6560502648353577,
      "learning_rate": 0.00036122185109363215,
      "loss": 3.4508,
      "step": 37050
    },
    {
      "epoch": 3.9931116133893014,
      "grad_norm": 0.5955866575241089,
      "learning_rate": 0.0003608986100635707,
      "loss": 3.4503,
      "step": 37100
    },
    {
      "epoch": 3.9984931654289095,
      "grad_norm": 0.6154381036758423,
      "learning_rate": 0.0003605753690335093,
      "loss": 3.4548,
      "step": 37150
    },
    {
      "epoch": 4.003874717468518,
      "grad_norm": 0.6376234292984009,
      "learning_rate": 0.0003602521280034478,
      "loss": 3.3711,
      "step": 37200
    },
    {
      "epoch": 4.009256269508126,
      "grad_norm": 0.6253707408905029,
      "learning_rate": 0.0003599288869733865,
      "loss": 3.3483,
      "step": 37250
    },
    {
      "epoch": 4.014637821547734,
      "grad_norm": 0.6334782838821411,
      "learning_rate": 0.00035960564594332507,
      "loss": 3.3408,
      "step": 37300
    },
    {
      "epoch": 4.020019373587343,
      "grad_norm": 0.6270228624343872,
      "learning_rate": 0.0003592824049132636,
      "loss": 3.3435,
      "step": 37350
    },
    {
      "epoch": 4.0254009256269505,
      "grad_norm": 0.6254666447639465,
      "learning_rate": 0.0003589591638832022,
      "loss": 3.3457,
      "step": 37400
    },
    {
      "epoch": 4.030782477666559,
      "grad_norm": 0.7249552607536316,
      "learning_rate": 0.0003586359228531408,
      "loss": 3.3514,
      "step": 37450
    },
    {
      "epoch": 4.036164029706168,
      "grad_norm": 0.6332011222839355,
      "learning_rate": 0.00035831268182307934,
      "loss": 3.3702,
      "step": 37500
    },
    {
      "epoch": 4.041545581745775,
      "grad_norm": 0.6144351959228516,
      "learning_rate": 0.000357989440793018,
      "loss": 3.3694,
      "step": 37550
    },
    {
      "epoch": 4.046927133785384,
      "grad_norm": 0.6586161851882935,
      "learning_rate": 0.0003576661997629566,
      "loss": 3.3739,
      "step": 37600
    },
    {
      "epoch": 4.0523086858249915,
      "grad_norm": 0.6490854024887085,
      "learning_rate": 0.0003573429587328951,
      "loss": 3.3556,
      "step": 37650
    },
    {
      "epoch": 4.0576902378646,
      "grad_norm": 0.6474526524543762,
      "learning_rate": 0.0003570197177028337,
      "loss": 3.3655,
      "step": 37700
    },
    {
      "epoch": 4.063071789904209,
      "grad_norm": 0.6664512157440186,
      "learning_rate": 0.00035669647667277226,
      "loss": 3.362,
      "step": 37750
    },
    {
      "epoch": 4.068453341943816,
      "grad_norm": 0.6286465525627136,
      "learning_rate": 0.00035637323564271085,
      "loss": 3.385,
      "step": 37800
    },
    {
      "epoch": 4.073834893983425,
      "grad_norm": 0.6096540093421936,
      "learning_rate": 0.0003560499946126495,
      "loss": 3.3604,
      "step": 37850
    },
    {
      "epoch": 4.079216446023033,
      "grad_norm": 0.6467686295509338,
      "learning_rate": 0.00035572675358258804,
      "loss": 3.3638,
      "step": 37900
    },
    {
      "epoch": 4.084597998062641,
      "grad_norm": 0.6298966407775879,
      "learning_rate": 0.00035540351255252664,
      "loss": 3.3726,
      "step": 37950
    },
    {
      "epoch": 4.08997955010225,
      "grad_norm": 0.7064566612243652,
      "learning_rate": 0.00035508027152246523,
      "loss": 3.376,
      "step": 38000
    },
    {
      "epoch": 4.08997955010225,
      "eval_accuracy": 0.37656059500929034,
      "eval_loss": 3.446718692779541,
      "eval_runtime": 184.5393,
      "eval_samples_per_second": 97.6,
      "eval_steps_per_second": 6.102,
      "step": 38000
    },
    {
      "epoch": 4.095361102141858,
      "grad_norm": 0.6466246843338013,
      "learning_rate": 0.0003547570304924038,
      "loss": 3.3837,
      "step": 38050
    },
    {
      "epoch": 4.100742654181466,
      "grad_norm": 0.6541309952735901,
      "learning_rate": 0.0003544337894623424,
      "loss": 3.3768,
      "step": 38100
    },
    {
      "epoch": 4.106124206221074,
      "grad_norm": 0.6219347715377808,
      "learning_rate": 0.000354110548432281,
      "loss": 3.361,
      "step": 38150
    },
    {
      "epoch": 4.111505758260682,
      "grad_norm": 0.6315196752548218,
      "learning_rate": 0.00035378730740221956,
      "loss": 3.3712,
      "step": 38200
    },
    {
      "epoch": 4.1168873103002905,
      "grad_norm": 0.6126587390899658,
      "learning_rate": 0.00035346406637215815,
      "loss": 3.3761,
      "step": 38250
    },
    {
      "epoch": 4.122268862339899,
      "grad_norm": 0.7233909368515015,
      "learning_rate": 0.0003531408253420967,
      "loss": 3.344,
      "step": 38300
    },
    {
      "epoch": 4.127650414379507,
      "grad_norm": 0.6340131163597107,
      "learning_rate": 0.0003528175843120353,
      "loss": 3.3787,
      "step": 38350
    },
    {
      "epoch": 4.133031966419115,
      "grad_norm": 0.6910709738731384,
      "learning_rate": 0.00035249434328197394,
      "loss": 3.3778,
      "step": 38400
    },
    {
      "epoch": 4.138413518458724,
      "grad_norm": 0.6195011138916016,
      "learning_rate": 0.0003521711022519125,
      "loss": 3.3878,
      "step": 38450
    },
    {
      "epoch": 4.1437950704983315,
      "grad_norm": 0.6645599007606506,
      "learning_rate": 0.00035184786122185107,
      "loss": 3.3773,
      "step": 38500
    },
    {
      "epoch": 4.14917662253794,
      "grad_norm": 0.6299995183944702,
      "learning_rate": 0.0003515310850123909,
      "loss": 3.3798,
      "step": 38550
    },
    {
      "epoch": 4.154558174577549,
      "grad_norm": 0.6655153632164001,
      "learning_rate": 0.00035120784398232947,
      "loss": 3.3805,
      "step": 38600
    },
    {
      "epoch": 4.159939726617156,
      "grad_norm": 0.626811683177948,
      "learning_rate": 0.000350884602952268,
      "loss": 3.3592,
      "step": 38650
    },
    {
      "epoch": 4.165321278656765,
      "grad_norm": 0.647689938545227,
      "learning_rate": 0.0003505613619222066,
      "loss": 3.383,
      "step": 38700
    },
    {
      "epoch": 4.1707028306963725,
      "grad_norm": 0.6576713919639587,
      "learning_rate": 0.0003502381208921452,
      "loss": 3.3792,
      "step": 38750
    },
    {
      "epoch": 4.176084382735981,
      "grad_norm": 0.6431198716163635,
      "learning_rate": 0.0003499148798620838,
      "loss": 3.3732,
      "step": 38800
    },
    {
      "epoch": 4.18146593477559,
      "grad_norm": 0.691774845123291,
      "learning_rate": 0.0003495916388320224,
      "loss": 3.3809,
      "step": 38850
    },
    {
      "epoch": 4.186847486815197,
      "grad_norm": 0.6469290852546692,
      "learning_rate": 0.000349268397801961,
      "loss": 3.3734,
      "step": 38900
    },
    {
      "epoch": 4.192229038854806,
      "grad_norm": 0.7078692317008972,
      "learning_rate": 0.0003489451567718995,
      "loss": 3.3864,
      "step": 38950
    },
    {
      "epoch": 4.197610590894414,
      "grad_norm": 0.6617368459701538,
      "learning_rate": 0.0003486219157418381,
      "loss": 3.3961,
      "step": 39000
    },
    {
      "epoch": 4.197610590894414,
      "eval_accuracy": 0.3772209872614268,
      "eval_loss": 3.4436137676239014,
      "eval_runtime": 184.7722,
      "eval_samples_per_second": 97.477,
      "eval_steps_per_second": 6.094,
      "step": 39000
    },
    {
      "epoch": 4.202992142934022,
      "grad_norm": 0.6935890316963196,
      "learning_rate": 0.00034829867471177677,
      "loss": 3.3808,
      "step": 39050
    },
    {
      "epoch": 4.208373694973631,
      "grad_norm": 0.6806594133377075,
      "learning_rate": 0.0003479754336817153,
      "loss": 3.3771,
      "step": 39100
    },
    {
      "epoch": 4.213755247013238,
      "grad_norm": 0.6456026434898376,
      "learning_rate": 0.0003476521926516539,
      "loss": 3.4124,
      "step": 39150
    },
    {
      "epoch": 4.219136799052847,
      "grad_norm": 0.5707885026931763,
      "learning_rate": 0.00034732895162159245,
      "loss": 3.3916,
      "step": 39200
    },
    {
      "epoch": 4.224518351092455,
      "grad_norm": 0.6426082849502563,
      "learning_rate": 0.00034700571059153104,
      "loss": 3.376,
      "step": 39250
    },
    {
      "epoch": 4.229899903132063,
      "grad_norm": 0.6489585638046265,
      "learning_rate": 0.00034668246956146963,
      "loss": 3.3809,
      "step": 39300
    },
    {
      "epoch": 4.2352814551716715,
      "grad_norm": 0.6257960796356201,
      "learning_rate": 0.00034635922853140823,
      "loss": 3.3873,
      "step": 39350
    },
    {
      "epoch": 4.24066300721128,
      "grad_norm": 0.698000967502594,
      "learning_rate": 0.0003460359875013468,
      "loss": 3.3771,
      "step": 39400
    },
    {
      "epoch": 4.246044559250888,
      "grad_norm": 0.6346425414085388,
      "learning_rate": 0.0003457127464712854,
      "loss": 3.3877,
      "step": 39450
    },
    {
      "epoch": 4.251426111290496,
      "grad_norm": 0.6701180338859558,
      "learning_rate": 0.00034538950544122396,
      "loss": 3.3697,
      "step": 39500
    },
    {
      "epoch": 4.256807663330104,
      "grad_norm": 0.6371496319770813,
      "learning_rate": 0.00034506626441116255,
      "loss": 3.3843,
      "step": 39550
    },
    {
      "epoch": 4.2621892153697125,
      "grad_norm": 0.6401668190956116,
      "learning_rate": 0.0003447430233811011,
      "loss": 3.3913,
      "step": 39600
    },
    {
      "epoch": 4.267570767409321,
      "grad_norm": 0.6551806926727295,
      "learning_rate": 0.00034441978235103974,
      "loss": 3.3918,
      "step": 39650
    },
    {
      "epoch": 4.272952319448929,
      "grad_norm": 0.6413207054138184,
      "learning_rate": 0.00034409654132097834,
      "loss": 3.3873,
      "step": 39700
    },
    {
      "epoch": 4.278333871488537,
      "grad_norm": 0.6472829580307007,
      "learning_rate": 0.0003437733002909169,
      "loss": 3.3974,
      "step": 39750
    },
    {
      "epoch": 4.283715423528146,
      "grad_norm": 0.6183416247367859,
      "learning_rate": 0.00034345005926085547,
      "loss": 3.3703,
      "step": 39800
    },
    {
      "epoch": 4.2890969755677535,
      "grad_norm": 0.715303361415863,
      "learning_rate": 0.00034312681823079407,
      "loss": 3.3877,
      "step": 39850
    },
    {
      "epoch": 4.294478527607362,
      "grad_norm": 0.6685398817062378,
      "learning_rate": 0.00034280357720073266,
      "loss": 3.3878,
      "step": 39900
    },
    {
      "epoch": 4.299860079646971,
      "grad_norm": 0.6901618242263794,
      "learning_rate": 0.00034248033617067126,
      "loss": 3.3789,
      "step": 39950
    },
    {
      "epoch": 4.305241631686578,
      "grad_norm": 0.6296851634979248,
      "learning_rate": 0.00034215709514060985,
      "loss": 3.3729,
      "step": 40000
    },
    {
      "epoch": 4.305241631686578,
      "eval_accuracy": 0.3777952177735263,
      "eval_loss": 3.437654972076416,
      "eval_runtime": 184.7949,
      "eval_samples_per_second": 97.465,
      "eval_steps_per_second": 6.093,
      "step": 40000
    },
    {
      "epoch": 4.310623183726187,
      "grad_norm": 0.7259793877601624,
      "learning_rate": 0.0003418338541105484,
      "loss": 3.3616,
      "step": 40050
    },
    {
      "epoch": 4.3160047357657945,
      "grad_norm": 0.6109891533851624,
      "learning_rate": 0.000341510613080487,
      "loss": 3.3857,
      "step": 40100
    },
    {
      "epoch": 4.321386287805403,
      "grad_norm": 0.6450995802879333,
      "learning_rate": 0.0003411873720504255,
      "loss": 3.389,
      "step": 40150
    },
    {
      "epoch": 4.326767839845012,
      "grad_norm": 0.6640621423721313,
      "learning_rate": 0.0003408641310203642,
      "loss": 3.3955,
      "step": 40200
    },
    {
      "epoch": 4.332149391884619,
      "grad_norm": 0.6404122710227966,
      "learning_rate": 0.00034054088999030277,
      "loss": 3.3787,
      "step": 40250
    },
    {
      "epoch": 4.337530943924228,
      "grad_norm": 0.6834564208984375,
      "learning_rate": 0.0003402176489602413,
      "loss": 3.3903,
      "step": 40300
    },
    {
      "epoch": 4.342912495963836,
      "grad_norm": 0.7080681324005127,
      "learning_rate": 0.0003398944079301799,
      "loss": 3.3965,
      "step": 40350
    },
    {
      "epoch": 4.348294048003444,
      "grad_norm": 0.6617289185523987,
      "learning_rate": 0.0003395711669001185,
      "loss": 3.4053,
      "step": 40400
    },
    {
      "epoch": 4.3536756000430525,
      "grad_norm": 0.6993653178215027,
      "learning_rate": 0.00033924792587005704,
      "loss": 3.3868,
      "step": 40450
    },
    {
      "epoch": 4.359057152082661,
      "grad_norm": 0.6167160868644714,
      "learning_rate": 0.0003389246848399957,
      "loss": 3.3719,
      "step": 40500
    },
    {
      "epoch": 4.364438704122269,
      "grad_norm": 0.669032096862793,
      "learning_rate": 0.0003386014438099343,
      "loss": 3.3802,
      "step": 40550
    },
    {
      "epoch": 4.369820256161877,
      "grad_norm": 0.619784414768219,
      "learning_rate": 0.0003382782027798728,
      "loss": 3.3882,
      "step": 40600
    },
    {
      "epoch": 4.375201808201485,
      "grad_norm": 0.6738349199295044,
      "learning_rate": 0.0003379549617498114,
      "loss": 3.3755,
      "step": 40650
    },
    {
      "epoch": 4.3805833602410935,
      "grad_norm": 0.6499971151351929,
      "learning_rate": 0.00033763172071974996,
      "loss": 3.3892,
      "step": 40700
    },
    {
      "epoch": 4.385964912280702,
      "grad_norm": 0.6084420084953308,
      "learning_rate": 0.00033730847968968855,
      "loss": 3.3835,
      "step": 40750
    },
    {
      "epoch": 4.39134646432031,
      "grad_norm": 0.6882473230361938,
      "learning_rate": 0.0003369852386596272,
      "loss": 3.3852,
      "step": 40800
    },
    {
      "epoch": 4.396728016359918,
      "grad_norm": 0.6895729303359985,
      "learning_rate": 0.00033666199762956574,
      "loss": 3.3923,
      "step": 40850
    },
    {
      "epoch": 4.402109568399527,
      "grad_norm": 0.6523504853248596,
      "learning_rate": 0.00033633875659950434,
      "loss": 3.3847,
      "step": 40900
    },
    {
      "epoch": 4.4074911204391345,
      "grad_norm": 0.619976282119751,
      "learning_rate": 0.0003360155155694429,
      "loss": 3.3709,
      "step": 40950
    },
    {
      "epoch": 4.412872672478743,
      "grad_norm": 0.6657698750495911,
      "learning_rate": 0.0003356922745393815,
      "loss": 3.3866,
      "step": 41000
    },
    {
      "epoch": 4.412872672478743,
      "eval_accuracy": 0.3780848863724397,
      "eval_loss": 3.4327142238616943,
      "eval_runtime": 184.5831,
      "eval_samples_per_second": 97.577,
      "eval_steps_per_second": 6.1,
      "step": 41000
    },
    {
      "epoch": 4.418254224518351,
      "grad_norm": 0.6433376669883728,
      "learning_rate": 0.0003353690335093201,
      "loss": 3.3842,
      "step": 41050
    },
    {
      "epoch": 4.423635776557959,
      "grad_norm": 0.676450252532959,
      "learning_rate": 0.0003350457924792587,
      "loss": 3.4097,
      "step": 41100
    },
    {
      "epoch": 4.429017328597568,
      "grad_norm": 0.6363053321838379,
      "learning_rate": 0.00033472255144919726,
      "loss": 3.389,
      "step": 41150
    },
    {
      "epoch": 4.4343988806371755,
      "grad_norm": 0.6782971620559692,
      "learning_rate": 0.00033439931041913585,
      "loss": 3.4028,
      "step": 41200
    },
    {
      "epoch": 4.439780432676784,
      "grad_norm": 0.6248264908790588,
      "learning_rate": 0.00033408253420967566,
      "loss": 3.3992,
      "step": 41250
    },
    {
      "epoch": 4.445161984716393,
      "grad_norm": 0.617743194103241,
      "learning_rate": 0.00033375929317961425,
      "loss": 3.3682,
      "step": 41300
    },
    {
      "epoch": 4.450543536756,
      "grad_norm": 0.7119070887565613,
      "learning_rate": 0.0003334360521495528,
      "loss": 3.3864,
      "step": 41350
    },
    {
      "epoch": 4.455925088795609,
      "grad_norm": 0.6565623879432678,
      "learning_rate": 0.0003331128111194914,
      "loss": 3.3814,
      "step": 41400
    },
    {
      "epoch": 4.461306640835216,
      "grad_norm": 0.7476305365562439,
      "learning_rate": 0.00033278957008943004,
      "loss": 3.384,
      "step": 41450
    },
    {
      "epoch": 4.466688192874825,
      "grad_norm": 0.6427007913589478,
      "learning_rate": 0.0003324663290593686,
      "loss": 3.3901,
      "step": 41500
    },
    {
      "epoch": 4.4720697449144335,
      "grad_norm": 0.6926642656326294,
      "learning_rate": 0.00033214308802930717,
      "loss": 3.3919,
      "step": 41550
    },
    {
      "epoch": 4.477451296954041,
      "grad_norm": 0.687822163105011,
      "learning_rate": 0.0003318198469992457,
      "loss": 3.3926,
      "step": 41600
    },
    {
      "epoch": 4.48283284899365,
      "grad_norm": 0.695188045501709,
      "learning_rate": 0.0003314966059691843,
      "loss": 3.3966,
      "step": 41650
    },
    {
      "epoch": 4.488214401033258,
      "grad_norm": 0.6738727688789368,
      "learning_rate": 0.00033117336493912296,
      "loss": 3.4002,
      "step": 41700
    },
    {
      "epoch": 4.493595953072866,
      "grad_norm": 0.6818880438804626,
      "learning_rate": 0.0003308501239090615,
      "loss": 3.3993,
      "step": 41750
    },
    {
      "epoch": 4.4989775051124745,
      "grad_norm": 0.6498254537582397,
      "learning_rate": 0.0003305268828790001,
      "loss": 3.4005,
      "step": 41800
    },
    {
      "epoch": 4.504359057152083,
      "grad_norm": 0.6302522420883179,
      "learning_rate": 0.0003302036418489387,
      "loss": 3.3798,
      "step": 41850
    },
    {
      "epoch": 4.509740609191691,
      "grad_norm": 0.6108139157295227,
      "learning_rate": 0.0003298804008188772,
      "loss": 3.36,
      "step": 41900
    },
    {
      "epoch": 4.515122161231299,
      "grad_norm": 0.6787891983985901,
      "learning_rate": 0.0003295571597888158,
      "loss": 3.3822,
      "step": 41950
    },
    {
      "epoch": 4.520503713270907,
      "grad_norm": 0.6845345497131348,
      "learning_rate": 0.00032923391875875447,
      "loss": 3.3773,
      "step": 42000
    },
    {
      "epoch": 4.520503713270907,
      "eval_accuracy": 0.3784473524061888,
      "eval_loss": 3.427448272705078,
      "eval_runtime": 184.8487,
      "eval_samples_per_second": 97.436,
      "eval_steps_per_second": 6.091,
      "step": 42000
    },
    {
      "epoch": 4.5258852653105155,
      "grad_norm": 0.6910591125488281,
      "learning_rate": 0.000328910677728693,
      "loss": 3.3944,
      "step": 42050
    },
    {
      "epoch": 4.531266817350124,
      "grad_norm": 0.6579834222793579,
      "learning_rate": 0.0003285874366986316,
      "loss": 3.3767,
      "step": 42100
    },
    {
      "epoch": 4.536648369389732,
      "grad_norm": 0.6646836996078491,
      "learning_rate": 0.00032826419566857015,
      "loss": 3.3925,
      "step": 42150
    },
    {
      "epoch": 4.54202992142934,
      "grad_norm": 0.6295957565307617,
      "learning_rate": 0.00032794095463850874,
      "loss": 3.3939,
      "step": 42200
    },
    {
      "epoch": 4.547411473468949,
      "grad_norm": 0.6931584477424622,
      "learning_rate": 0.0003276177136084473,
      "loss": 3.3813,
      "step": 42250
    },
    {
      "epoch": 4.5527930255085565,
      "grad_norm": 0.728244423866272,
      "learning_rate": 0.00032729447257838593,
      "loss": 3.3823,
      "step": 42300
    },
    {
      "epoch": 4.558174577548165,
      "grad_norm": 0.6908766627311707,
      "learning_rate": 0.0003269712315483245,
      "loss": 3.3996,
      "step": 42350
    },
    {
      "epoch": 4.563556129587774,
      "grad_norm": 0.6593059301376343,
      "learning_rate": 0.00032664799051826306,
      "loss": 3.3678,
      "step": 42400
    },
    {
      "epoch": 4.568937681627381,
      "grad_norm": 0.7266299724578857,
      "learning_rate": 0.00032632474948820166,
      "loss": 3.3908,
      "step": 42450
    },
    {
      "epoch": 4.57431923366699,
      "grad_norm": 0.6872548460960388,
      "learning_rate": 0.00032600150845814025,
      "loss": 3.3937,
      "step": 42500
    },
    {
      "epoch": 4.579700785706597,
      "grad_norm": 0.7212083339691162,
      "learning_rate": 0.0003256782674280788,
      "loss": 3.3884,
      "step": 42550
    },
    {
      "epoch": 4.585082337746206,
      "grad_norm": 0.6755133867263794,
      "learning_rate": 0.00032535502639801744,
      "loss": 3.3797,
      "step": 42600
    },
    {
      "epoch": 4.5904638897858145,
      "grad_norm": 0.634061336517334,
      "learning_rate": 0.00032503178536795604,
      "loss": 3.3579,
      "step": 42650
    },
    {
      "epoch": 4.595845441825422,
      "grad_norm": 0.6569266319274902,
      "learning_rate": 0.0003247085443378946,
      "loss": 3.3904,
      "step": 42700
    },
    {
      "epoch": 4.601226993865031,
      "grad_norm": 0.6839199066162109,
      "learning_rate": 0.0003243853033078332,
      "loss": 3.3873,
      "step": 42750
    },
    {
      "epoch": 4.606608545904638,
      "grad_norm": 0.6381998062133789,
      "learning_rate": 0.0003240620622777717,
      "loss": 3.3958,
      "step": 42800
    },
    {
      "epoch": 4.611990097944247,
      "grad_norm": 0.6814762949943542,
      "learning_rate": 0.00032373882124771036,
      "loss": 3.3978,
      "step": 42850
    },
    {
      "epoch": 4.6173716499838555,
      "grad_norm": 0.6227290630340576,
      "learning_rate": 0.00032341558021764896,
      "loss": 3.3834,
      "step": 42900
    },
    {
      "epoch": 4.622753202023463,
      "grad_norm": 0.644250750541687,
      "learning_rate": 0.0003230923391875875,
      "loss": 3.3759,
      "step": 42950
    },
    {
      "epoch": 4.628134754063072,
      "grad_norm": 0.6935346722602844,
      "learning_rate": 0.0003227690981575261,
      "loss": 3.4146,
      "step": 43000
    },
    {
      "epoch": 4.628134754063072,
      "eval_accuracy": 0.37911926186443357,
      "eval_loss": 3.4232747554779053,
      "eval_runtime": 184.608,
      "eval_samples_per_second": 97.563,
      "eval_steps_per_second": 6.099,
      "step": 43000
    },
    {
      "epoch": 4.63351630610268,
      "grad_norm": 0.6674992442131042,
      "learning_rate": 0.0003224458571274647,
      "loss": 3.3878,
      "step": 43050
    },
    {
      "epoch": 4.638897858142288,
      "grad_norm": 0.6759820580482483,
      "learning_rate": 0.00032212261609740323,
      "loss": 3.3895,
      "step": 43100
    },
    {
      "epoch": 4.6442794101818965,
      "grad_norm": 0.687537670135498,
      "learning_rate": 0.0003217993750673419,
      "loss": 3.3907,
      "step": 43150
    },
    {
      "epoch": 4.649660962221505,
      "grad_norm": 0.6385945081710815,
      "learning_rate": 0.00032147613403728047,
      "loss": 3.3825,
      "step": 43200
    },
    {
      "epoch": 4.655042514261113,
      "grad_norm": 0.7066798210144043,
      "learning_rate": 0.000321152893007219,
      "loss": 3.3945,
      "step": 43250
    },
    {
      "epoch": 4.660424066300721,
      "grad_norm": 0.7198359966278076,
      "learning_rate": 0.0003208296519771576,
      "loss": 3.4057,
      "step": 43300
    },
    {
      "epoch": 4.665805618340329,
      "grad_norm": 0.6627112627029419,
      "learning_rate": 0.00032050641094709615,
      "loss": 3.3895,
      "step": 43350
    },
    {
      "epoch": 4.6711871703799375,
      "grad_norm": 0.6638057827949524,
      "learning_rate": 0.00032018316991703474,
      "loss": 3.3827,
      "step": 43400
    },
    {
      "epoch": 4.676568722419546,
      "grad_norm": 0.641120970249176,
      "learning_rate": 0.0003198599288869734,
      "loss": 3.3974,
      "step": 43450
    },
    {
      "epoch": 4.681950274459154,
      "grad_norm": 0.6419323682785034,
      "learning_rate": 0.00031953668785691193,
      "loss": 3.3961,
      "step": 43500
    },
    {
      "epoch": 4.687331826498762,
      "grad_norm": 0.7281374335289001,
      "learning_rate": 0.0003192134468268505,
      "loss": 3.3956,
      "step": 43550
    },
    {
      "epoch": 4.692713378538371,
      "grad_norm": 0.6420050859451294,
      "learning_rate": 0.0003188902057967891,
      "loss": 3.3683,
      "step": 43600
    },
    {
      "epoch": 4.6980949305779784,
      "grad_norm": 0.6763567328453064,
      "learning_rate": 0.00031856696476672766,
      "loss": 3.3961,
      "step": 43650
    },
    {
      "epoch": 4.703476482617587,
      "grad_norm": 0.7230156064033508,
      "learning_rate": 0.0003182437237366663,
      "loss": 3.4111,
      "step": 43700
    },
    {
      "epoch": 4.7088580346571955,
      "grad_norm": 0.7006603479385376,
      "learning_rate": 0.0003179204827066049,
      "loss": 3.4003,
      "step": 43750
    },
    {
      "epoch": 4.714239586696803,
      "grad_norm": 0.6924633383750916,
      "learning_rate": 0.00031759724167654344,
      "loss": 3.3832,
      "step": 43800
    },
    {
      "epoch": 4.719621138736412,
      "grad_norm": 0.7065801024436951,
      "learning_rate": 0.00031727400064648204,
      "loss": 3.3894,
      "step": 43850
    },
    {
      "epoch": 4.725002690776019,
      "grad_norm": 0.6304015517234802,
      "learning_rate": 0.0003169507596164206,
      "loss": 3.3659,
      "step": 43900
    },
    {
      "epoch": 4.730384242815628,
      "grad_norm": 0.6763742566108704,
      "learning_rate": 0.0003166275185863592,
      "loss": 3.3802,
      "step": 43950
    },
    {
      "epoch": 4.7357657948552365,
      "grad_norm": 0.6544877886772156,
      "learning_rate": 0.0003163042775562978,
      "loss": 3.3892,
      "step": 44000
    },
    {
      "epoch": 4.7357657948552365,
      "eval_accuracy": 0.3791140465258185,
      "eval_loss": 3.4197874069213867,
      "eval_runtime": 184.7556,
      "eval_samples_per_second": 97.486,
      "eval_steps_per_second": 6.095,
      "step": 44000
    },
    {
      "epoch": 4.741147346894844,
      "grad_norm": 0.6808891296386719,
      "learning_rate": 0.00031598103652623636,
      "loss": 3.3747,
      "step": 44050
    },
    {
      "epoch": 4.746528898934453,
      "grad_norm": 0.6510489583015442,
      "learning_rate": 0.00031565779549617496,
      "loss": 3.3851,
      "step": 44100
    },
    {
      "epoch": 4.751910450974061,
      "grad_norm": 0.659172773361206,
      "learning_rate": 0.00031534101928671476,
      "loss": 3.3789,
      "step": 44150
    },
    {
      "epoch": 4.757292003013669,
      "grad_norm": 0.7054234743118286,
      "learning_rate": 0.00031501777825665336,
      "loss": 3.3755,
      "step": 44200
    },
    {
      "epoch": 4.7626735550532775,
      "grad_norm": 0.7176251411437988,
      "learning_rate": 0.0003146945372265919,
      "loss": 3.3863,
      "step": 44250
    },
    {
      "epoch": 4.768055107092886,
      "grad_norm": 0.751799464225769,
      "learning_rate": 0.0003143712961965305,
      "loss": 3.3894,
      "step": 44300
    },
    {
      "epoch": 4.773436659132494,
      "grad_norm": 0.6770383715629578,
      "learning_rate": 0.0003140480551664691,
      "loss": 3.3808,
      "step": 44350
    },
    {
      "epoch": 4.778818211172102,
      "grad_norm": 0.6924847960472107,
      "learning_rate": 0.0003137248141364077,
      "loss": 3.3874,
      "step": 44400
    },
    {
      "epoch": 4.78419976321171,
      "grad_norm": 0.633284330368042,
      "learning_rate": 0.0003134015731063463,
      "loss": 3.3971,
      "step": 44450
    },
    {
      "epoch": 4.7895813152513185,
      "grad_norm": 0.6829878091812134,
      "learning_rate": 0.0003130783320762849,
      "loss": 3.3811,
      "step": 44500
    },
    {
      "epoch": 4.794962867290927,
      "grad_norm": 0.6898048520088196,
      "learning_rate": 0.0003127550910462234,
      "loss": 3.3897,
      "step": 44550
    },
    {
      "epoch": 4.800344419330535,
      "grad_norm": 0.6750717163085938,
      "learning_rate": 0.000312431850016162,
      "loss": 3.3971,
      "step": 44600
    },
    {
      "epoch": 4.805725971370143,
      "grad_norm": 0.6479890942573547,
      "learning_rate": 0.00031210860898610066,
      "loss": 3.406,
      "step": 44650
    },
    {
      "epoch": 4.811107523409751,
      "grad_norm": 0.7591122984886169,
      "learning_rate": 0.0003117853679560392,
      "loss": 3.3913,
      "step": 44700
    },
    {
      "epoch": 4.8164890754493594,
      "grad_norm": 0.6572939157485962,
      "learning_rate": 0.0003114621269259778,
      "loss": 3.38,
      "step": 44750
    },
    {
      "epoch": 4.821870627488968,
      "grad_norm": 0.6863782405853271,
      "learning_rate": 0.00031113888589591633,
      "loss": 3.3788,
      "step": 44800
    },
    {
      "epoch": 4.827252179528576,
      "grad_norm": 0.6886435747146606,
      "learning_rate": 0.00031081564486585493,
      "loss": 3.3875,
      "step": 44850
    },
    {
      "epoch": 4.832633731568184,
      "grad_norm": 0.7485862970352173,
      "learning_rate": 0.0003104924038357935,
      "loss": 3.3789,
      "step": 44900
    },
    {
      "epoch": 4.838015283607793,
      "grad_norm": 0.666218638420105,
      "learning_rate": 0.0003101691628057321,
      "loss": 3.3945,
      "step": 44950
    },
    {
      "epoch": 4.8433968356474,
      "grad_norm": 0.6517705917358398,
      "learning_rate": 0.0003098459217756707,
      "loss": 3.401,
      "step": 45000
    },
    {
      "epoch": 4.8433968356474,
      "eval_accuracy": 0.38019872830487045,
      "eval_loss": 3.4125325679779053,
      "eval_runtime": 184.9758,
      "eval_samples_per_second": 97.369,
      "eval_steps_per_second": 6.087,
      "step": 45000
    },
    {
      "epoch": 4.848778387687009,
      "grad_norm": 0.6660953760147095,
      "learning_rate": 0.0003095226807456093,
      "loss": 3.3751,
      "step": 45050
    },
    {
      "epoch": 4.8541599397266175,
      "grad_norm": 0.6609524488449097,
      "learning_rate": 0.00030919943971554785,
      "loss": 3.3946,
      "step": 45100
    },
    {
      "epoch": 4.859541491766225,
      "grad_norm": 0.6775374412536621,
      "learning_rate": 0.00030887619868548644,
      "loss": 3.3883,
      "step": 45150
    },
    {
      "epoch": 4.864923043805834,
      "grad_norm": 0.664930522441864,
      "learning_rate": 0.000308552957655425,
      "loss": 3.3975,
      "step": 45200
    },
    {
      "epoch": 4.870304595845441,
      "grad_norm": 0.6616421341896057,
      "learning_rate": 0.00030822971662536363,
      "loss": 3.3899,
      "step": 45250
    },
    {
      "epoch": 4.87568614788505,
      "grad_norm": 0.7150900959968567,
      "learning_rate": 0.0003079064755953022,
      "loss": 3.3785,
      "step": 45300
    },
    {
      "epoch": 4.8810676999246585,
      "grad_norm": 0.6499496698379517,
      "learning_rate": 0.00030758323456524077,
      "loss": 3.3818,
      "step": 45350
    },
    {
      "epoch": 4.886449251964266,
      "grad_norm": 0.658064603805542,
      "learning_rate": 0.00030725999353517936,
      "loss": 3.3796,
      "step": 45400
    },
    {
      "epoch": 4.891830804003875,
      "grad_norm": 0.6768323183059692,
      "learning_rate": 0.00030693675250511795,
      "loss": 3.3858,
      "step": 45450
    },
    {
      "epoch": 4.897212356043483,
      "grad_norm": 0.679030179977417,
      "learning_rate": 0.00030661351147505655,
      "loss": 3.3737,
      "step": 45500
    },
    {
      "epoch": 4.902593908083091,
      "grad_norm": 0.6560843586921692,
      "learning_rate": 0.00030629027044499514,
      "loss": 3.3828,
      "step": 45550
    },
    {
      "epoch": 4.9079754601226995,
      "grad_norm": 0.6728476881980896,
      "learning_rate": 0.00030596702941493374,
      "loss": 3.3804,
      "step": 45600
    },
    {
      "epoch": 4.913357012162308,
      "grad_norm": 0.736443817615509,
      "learning_rate": 0.0003056437883848723,
      "loss": 3.3771,
      "step": 45650
    },
    {
      "epoch": 4.918738564201916,
      "grad_norm": 0.6769527792930603,
      "learning_rate": 0.0003053205473548109,
      "loss": 3.4014,
      "step": 45700
    },
    {
      "epoch": 4.924120116241524,
      "grad_norm": 0.6532332301139832,
      "learning_rate": 0.0003049973063247494,
      "loss": 3.3797,
      "step": 45750
    },
    {
      "epoch": 4.929501668281132,
      "grad_norm": 0.6099948287010193,
      "learning_rate": 0.00030467406529468806,
      "loss": 3.3837,
      "step": 45800
    },
    {
      "epoch": 4.9348832203207404,
      "grad_norm": 0.6482587456703186,
      "learning_rate": 0.00030435082426462666,
      "loss": 3.3776,
      "step": 45850
    },
    {
      "epoch": 4.940264772360349,
      "grad_norm": 0.6524862051010132,
      "learning_rate": 0.0003040275832345652,
      "loss": 3.3893,
      "step": 45900
    },
    {
      "epoch": 4.945646324399957,
      "grad_norm": 0.6760078072547913,
      "learning_rate": 0.0003037043422045038,
      "loss": 3.386,
      "step": 45950
    },
    {
      "epoch": 4.951027876439565,
      "grad_norm": 0.7402188777923584,
      "learning_rate": 0.0003033811011744424,
      "loss": 3.3919,
      "step": 46000
    },
    {
      "epoch": 4.951027876439565,
      "eval_accuracy": 0.3805924863703101,
      "eval_loss": 3.4080684185028076,
      "eval_runtime": 184.4763,
      "eval_samples_per_second": 97.633,
      "eval_steps_per_second": 6.104,
      "step": 46000
    },
    {
      "epoch": 4.956409428479174,
      "grad_norm": 0.7083448767662048,
      "learning_rate": 0.00030305786014438093,
      "loss": 3.3806,
      "step": 46050
    },
    {
      "epoch": 4.961790980518781,
      "grad_norm": 0.7174310684204102,
      "learning_rate": 0.0003027346191143196,
      "loss": 3.37,
      "step": 46100
    },
    {
      "epoch": 4.96717253255839,
      "grad_norm": 0.6411959528923035,
      "learning_rate": 0.00030241137808425817,
      "loss": 3.3813,
      "step": 46150
    },
    {
      "epoch": 4.9725540845979985,
      "grad_norm": 0.6771125197410583,
      "learning_rate": 0.0003020881370541967,
      "loss": 3.3759,
      "step": 46200
    },
    {
      "epoch": 4.977935636637606,
      "grad_norm": 0.6481576561927795,
      "learning_rate": 0.0003017648960241353,
      "loss": 3.3786,
      "step": 46250
    },
    {
      "epoch": 4.983317188677215,
      "grad_norm": 0.6613046526908875,
      "learning_rate": 0.00030144165499407385,
      "loss": 3.4187,
      "step": 46300
    },
    {
      "epoch": 4.988698740716822,
      "grad_norm": 0.6821366548538208,
      "learning_rate": 0.00030111841396401244,
      "loss": 3.3862,
      "step": 46350
    },
    {
      "epoch": 4.994080292756431,
      "grad_norm": 0.7078256607055664,
      "learning_rate": 0.0003007951729339511,
      "loss": 3.3917,
      "step": 46400
    },
    {
      "epoch": 4.9994618447960395,
      "grad_norm": 0.6903404593467712,
      "learning_rate": 0.00030047193190388963,
      "loss": 3.3968,
      "step": 46450
    },
    {
      "epoch": 5.004843396835647,
      "grad_norm": 0.6922981142997742,
      "learning_rate": 0.0003001486908738282,
      "loss": 3.3085,
      "step": 46500
    },
    {
      "epoch": 5.010224948875256,
      "grad_norm": 0.75609290599823,
      "learning_rate": 0.0002998254498437668,
      "loss": 3.2916,
      "step": 46550
    },
    {
      "epoch": 5.015606500914864,
      "grad_norm": 0.7146428823471069,
      "learning_rate": 0.0002995022088137054,
      "loss": 3.3115,
      "step": 46600
    },
    {
      "epoch": 5.020988052954472,
      "grad_norm": 0.7553977966308594,
      "learning_rate": 0.00029917896778364396,
      "loss": 3.3028,
      "step": 46650
    },
    {
      "epoch": 5.0263696049940805,
      "grad_norm": 0.6598968505859375,
      "learning_rate": 0.00029885572675358255,
      "loss": 3.2995,
      "step": 46700
    },
    {
      "epoch": 5.031751157033688,
      "grad_norm": 0.6931416988372803,
      "learning_rate": 0.00029853248572352114,
      "loss": 3.2986,
      "step": 46750
    },
    {
      "epoch": 5.037132709073297,
      "grad_norm": 0.6862365007400513,
      "learning_rate": 0.00029820924469345974,
      "loss": 3.3076,
      "step": 46800
    },
    {
      "epoch": 5.042514261112905,
      "grad_norm": 0.674670398235321,
      "learning_rate": 0.0002978860036633983,
      "loss": 3.282,
      "step": 46850
    },
    {
      "epoch": 5.047895813152513,
      "grad_norm": 0.6725071668624878,
      "learning_rate": 0.00029756276263333693,
      "loss": 3.3175,
      "step": 46900
    },
    {
      "epoch": 5.0532773651921215,
      "grad_norm": 0.715003252029419,
      "learning_rate": 0.00029723952160327547,
      "loss": 3.3009,
      "step": 46950
    },
    {
      "epoch": 5.05865891723173,
      "grad_norm": 0.6437234878540039,
      "learning_rate": 0.00029691628057321406,
      "loss": 3.2957,
      "step": 47000
    },
    {
      "epoch": 5.05865891723173,
      "eval_accuracy": 0.3807873009981615,
      "eval_loss": 3.412716865539551,
      "eval_runtime": 184.8748,
      "eval_samples_per_second": 97.423,
      "eval_steps_per_second": 6.091,
      "step": 47000
    },
    {
      "epoch": 5.064040469271338,
      "grad_norm": 0.6628379225730896,
      "learning_rate": 0.00029659303954315266,
      "loss": 3.3054,
      "step": 47050
    },
    {
      "epoch": 5.069422021310946,
      "grad_norm": 0.6656396985054016,
      "learning_rate": 0.00029626979851309125,
      "loss": 3.3118,
      "step": 47100
    },
    {
      "epoch": 5.074803573350554,
      "grad_norm": 0.6756969094276428,
      "learning_rate": 0.00029594655748302985,
      "loss": 3.2969,
      "step": 47150
    },
    {
      "epoch": 5.080185125390162,
      "grad_norm": 0.6976021528244019,
      "learning_rate": 0.0002956233164529684,
      "loss": 3.2881,
      "step": 47200
    },
    {
      "epoch": 5.085566677429771,
      "grad_norm": 0.7710605263710022,
      "learning_rate": 0.000295300075422907,
      "loss": 3.3236,
      "step": 47250
    },
    {
      "epoch": 5.090948229469379,
      "grad_norm": 0.6808165907859802,
      "learning_rate": 0.0002949768343928456,
      "loss": 3.3053,
      "step": 47300
    },
    {
      "epoch": 5.096329781508987,
      "grad_norm": 0.6778217554092407,
      "learning_rate": 0.00029465359336278417,
      "loss": 3.3169,
      "step": 47350
    },
    {
      "epoch": 5.101711333548596,
      "grad_norm": 0.7051342725753784,
      "learning_rate": 0.0002943303523327227,
      "loss": 3.3148,
      "step": 47400
    },
    {
      "epoch": 5.107092885588203,
      "grad_norm": 0.6564488410949707,
      "learning_rate": 0.00029400711130266136,
      "loss": 3.2898,
      "step": 47450
    },
    {
      "epoch": 5.112474437627812,
      "grad_norm": 0.6671553254127502,
      "learning_rate": 0.0002936838702725999,
      "loss": 3.3092,
      "step": 47500
    },
    {
      "epoch": 5.1178559896674205,
      "grad_norm": 0.7125695943832397,
      "learning_rate": 0.0002933606292425385,
      "loss": 3.3329,
      "step": 47550
    },
    {
      "epoch": 5.123237541707028,
      "grad_norm": 0.7285773754119873,
      "learning_rate": 0.0002930373882124771,
      "loss": 3.3005,
      "step": 47600
    },
    {
      "epoch": 5.128619093746637,
      "grad_norm": 0.7182151675224304,
      "learning_rate": 0.0002927141471824157,
      "loss": 3.3026,
      "step": 47650
    },
    {
      "epoch": 5.134000645786244,
      "grad_norm": 0.6936147212982178,
      "learning_rate": 0.0002923909061523542,
      "loss": 3.3223,
      "step": 47700
    },
    {
      "epoch": 5.139382197825853,
      "grad_norm": 0.7076971530914307,
      "learning_rate": 0.0002920676651222928,
      "loss": 3.3132,
      "step": 47750
    },
    {
      "epoch": 5.1447637498654615,
      "grad_norm": 0.7523118257522583,
      "learning_rate": 0.0002917444240922314,
      "loss": 3.331,
      "step": 47800
    },
    {
      "epoch": 5.150145301905069,
      "grad_norm": 0.6461667418479919,
      "learning_rate": 0.00029142118306216996,
      "loss": 3.3323,
      "step": 47850
    },
    {
      "epoch": 5.155526853944678,
      "grad_norm": 0.6895806193351746,
      "learning_rate": 0.0002910979420321086,
      "loss": 3.3141,
      "step": 47900
    },
    {
      "epoch": 5.160908405984286,
      "grad_norm": 0.715032696723938,
      "learning_rate": 0.00029077470100204715,
      "loss": 3.3178,
      "step": 47950
    },
    {
      "epoch": 5.166289958023894,
      "grad_norm": 0.7430562376976013,
      "learning_rate": 0.00029045145997198574,
      "loss": 3.3146,
      "step": 48000
    },
    {
      "epoch": 5.166289958023894,
      "eval_accuracy": 0.38082771987242847,
      "eval_loss": 3.4105722904205322,
      "eval_runtime": 184.8541,
      "eval_samples_per_second": 97.434,
      "eval_steps_per_second": 6.091,
      "step": 48000
    },
    {
      "epoch": 5.1716715100635025,
      "grad_norm": 0.6621975302696228,
      "learning_rate": 0.00029012821894192433,
      "loss": 3.3294,
      "step": 48050
    },
    {
      "epoch": 5.17705306210311,
      "grad_norm": 0.6793010830879211,
      "learning_rate": 0.00028980497791186293,
      "loss": 3.3091,
      "step": 48100
    },
    {
      "epoch": 5.182434614142719,
      "grad_norm": 0.7010188698768616,
      "learning_rate": 0.00028948820170240274,
      "loss": 3.3182,
      "step": 48150
    },
    {
      "epoch": 5.187816166182327,
      "grad_norm": 0.6465266942977905,
      "learning_rate": 0.00028916496067234133,
      "loss": 3.3094,
      "step": 48200
    },
    {
      "epoch": 5.193197718221935,
      "grad_norm": 0.7279083132743835,
      "learning_rate": 0.0002888417196422799,
      "loss": 3.3092,
      "step": 48250
    },
    {
      "epoch": 5.198579270261543,
      "grad_norm": 0.6986241936683655,
      "learning_rate": 0.00028851847861221847,
      "loss": 3.2993,
      "step": 48300
    },
    {
      "epoch": 5.203960822301152,
      "grad_norm": 0.6948270797729492,
      "learning_rate": 0.00028819523758215706,
      "loss": 3.3012,
      "step": 48350
    },
    {
      "epoch": 5.20934237434076,
      "grad_norm": 0.6764682531356812,
      "learning_rate": 0.00028787199655209566,
      "loss": 3.3296,
      "step": 48400
    },
    {
      "epoch": 5.214723926380368,
      "grad_norm": 0.7321643233299255,
      "learning_rate": 0.00028754875552203425,
      "loss": 3.3373,
      "step": 48450
    },
    {
      "epoch": 5.220105478419977,
      "grad_norm": 0.666074275970459,
      "learning_rate": 0.0002872255144919728,
      "loss": 3.315,
      "step": 48500
    },
    {
      "epoch": 5.225487030459584,
      "grad_norm": 0.7783872485160828,
      "learning_rate": 0.00028690227346191144,
      "loss": 3.3212,
      "step": 48550
    },
    {
      "epoch": 5.230868582499193,
      "grad_norm": 0.726060152053833,
      "learning_rate": 0.00028657903243185,
      "loss": 3.3096,
      "step": 48600
    },
    {
      "epoch": 5.236250134538801,
      "grad_norm": 0.6934331655502319,
      "learning_rate": 0.0002862557914017886,
      "loss": 3.3258,
      "step": 48650
    },
    {
      "epoch": 5.241631686578409,
      "grad_norm": 0.684074878692627,
      "learning_rate": 0.00028593255037172717,
      "loss": 3.3216,
      "step": 48700
    },
    {
      "epoch": 5.247013238618018,
      "grad_norm": 0.6767908334732056,
      "learning_rate": 0.00028560930934166576,
      "loss": 3.3207,
      "step": 48750
    },
    {
      "epoch": 5.252394790657625,
      "grad_norm": 0.691230058670044,
      "learning_rate": 0.00028528606831160436,
      "loss": 3.3434,
      "step": 48800
    },
    {
      "epoch": 5.257776342697234,
      "grad_norm": 0.7426613569259644,
      "learning_rate": 0.0002849628272815429,
      "loss": 3.3263,
      "step": 48850
    },
    {
      "epoch": 5.2631578947368425,
      "grad_norm": 0.6871113777160645,
      "learning_rate": 0.0002846395862514815,
      "loss": 3.3332,
      "step": 48900
    },
    {
      "epoch": 5.26853944677645,
      "grad_norm": 0.6987447142601013,
      "learning_rate": 0.0002843228100420213,
      "loss": 3.3349,
      "step": 48950
    },
    {
      "epoch": 5.273920998816059,
      "grad_norm": 0.6851251125335693,
      "learning_rate": 0.0002839995690119599,
      "loss": 3.3387,
      "step": 49000
    },
    {
      "epoch": 5.273920998816059,
      "eval_accuracy": 0.3813237203053016,
      "eval_loss": 3.4056828022003174,
      "eval_runtime": 184.6477,
      "eval_samples_per_second": 97.543,
      "eval_steps_per_second": 6.098,
      "step": 49000
    },
    {
      "epoch": 5.279302550855666,
      "grad_norm": 0.6929810047149658,
      "learning_rate": 0.0002836763279818985,
      "loss": 3.3249,
      "step": 49050
    },
    {
      "epoch": 5.284684102895275,
      "grad_norm": 0.7261629104614258,
      "learning_rate": 0.0002833530869518371,
      "loss": 3.3331,
      "step": 49100
    },
    {
      "epoch": 5.2900656549348835,
      "grad_norm": 0.6819291114807129,
      "learning_rate": 0.0002830298459217756,
      "loss": 3.326,
      "step": 49150
    },
    {
      "epoch": 5.295447206974491,
      "grad_norm": 0.7193529009819031,
      "learning_rate": 0.0002827066048917142,
      "loss": 3.3126,
      "step": 49200
    },
    {
      "epoch": 5.3008287590141,
      "grad_norm": 0.7201562523841858,
      "learning_rate": 0.0002823833638616528,
      "loss": 3.3113,
      "step": 49250
    },
    {
      "epoch": 5.306210311053708,
      "grad_norm": 0.6952700614929199,
      "learning_rate": 0.0002820601228315914,
      "loss": 3.3216,
      "step": 49300
    },
    {
      "epoch": 5.311591863093316,
      "grad_norm": 0.6874309778213501,
      "learning_rate": 0.00028173688180153,
      "loss": 3.3205,
      "step": 49350
    },
    {
      "epoch": 5.316973415132924,
      "grad_norm": 0.6886963844299316,
      "learning_rate": 0.00028141364077146854,
      "loss": 3.3213,
      "step": 49400
    },
    {
      "epoch": 5.322354967172533,
      "grad_norm": 0.7013046741485596,
      "learning_rate": 0.00028109039974140714,
      "loss": 3.344,
      "step": 49450
    },
    {
      "epoch": 5.327736519212141,
      "grad_norm": 0.73094242811203,
      "learning_rate": 0.00028076715871134573,
      "loss": 3.3239,
      "step": 49500
    },
    {
      "epoch": 5.333118071251749,
      "grad_norm": 0.7266910076141357,
      "learning_rate": 0.00028044391768128433,
      "loss": 3.3593,
      "step": 49550
    },
    {
      "epoch": 5.338499623291357,
      "grad_norm": 0.6710968017578125,
      "learning_rate": 0.0002801206766512229,
      "loss": 3.3211,
      "step": 49600
    },
    {
      "epoch": 5.343881175330965,
      "grad_norm": 0.6896623373031616,
      "learning_rate": 0.0002797974356211615,
      "loss": 3.3291,
      "step": 49650
    },
    {
      "epoch": 5.349262727370574,
      "grad_norm": 0.7244205474853516,
      "learning_rate": 0.00027947419459110006,
      "loss": 3.3124,
      "step": 49700
    },
    {
      "epoch": 5.354644279410182,
      "grad_norm": 0.7222291827201843,
      "learning_rate": 0.00027915095356103865,
      "loss": 3.3257,
      "step": 49750
    },
    {
      "epoch": 5.36002583144979,
      "grad_norm": 0.6412253379821777,
      "learning_rate": 0.00027882771253097725,
      "loss": 3.3272,
      "step": 49800
    },
    {
      "epoch": 5.365407383489399,
      "grad_norm": 0.699478268623352,
      "learning_rate": 0.00027850447150091584,
      "loss": 3.335,
      "step": 49850
    },
    {
      "epoch": 5.370788935529006,
      "grad_norm": 0.6737022399902344,
      "learning_rate": 0.00027818123047085444,
      "loss": 3.3194,
      "step": 49900
    },
    {
      "epoch": 5.376170487568615,
      "grad_norm": 0.6805015802383423,
      "learning_rate": 0.000277857989440793,
      "loss": 3.3292,
      "step": 49950
    },
    {
      "epoch": 5.3815520396082235,
      "grad_norm": 0.7129510641098022,
      "learning_rate": 0.00027753474841073157,
      "loss": 3.3388,
      "step": 50000
    },
    {
      "epoch": 5.3815520396082235,
      "eval_accuracy": 0.3819017536684746,
      "eval_loss": 3.400902032852173,
      "eval_runtime": 184.6882,
      "eval_samples_per_second": 97.521,
      "eval_steps_per_second": 6.097,
      "step": 50000
    },
    {
      "epoch": 5.386933591647831,
      "grad_norm": 0.7065452933311462,
      "learning_rate": 0.00027721150738067017,
      "loss": 3.3545,
      "step": 50050
    },
    {
      "epoch": 5.39231514368744,
      "grad_norm": 0.714790403842926,
      "learning_rate": 0.00027688826635060876,
      "loss": 3.3329,
      "step": 50100
    },
    {
      "epoch": 5.397696695727047,
      "grad_norm": 0.7263872027397156,
      "learning_rate": 0.0002765650253205473,
      "loss": 3.3431,
      "step": 50150
    },
    {
      "epoch": 5.403078247766656,
      "grad_norm": 0.6966455578804016,
      "learning_rate": 0.00027624178429048595,
      "loss": 3.3286,
      "step": 50200
    },
    {
      "epoch": 5.4084597998062645,
      "grad_norm": 0.716153621673584,
      "learning_rate": 0.0002759185432604245,
      "loss": 3.3326,
      "step": 50250
    },
    {
      "epoch": 5.413841351845872,
      "grad_norm": 0.7463889718055725,
      "learning_rate": 0.0002755953022303631,
      "loss": 3.3212,
      "step": 50300
    },
    {
      "epoch": 5.419222903885481,
      "grad_norm": 0.6929644346237183,
      "learning_rate": 0.0002752720612003017,
      "loss": 3.3268,
      "step": 50350
    },
    {
      "epoch": 5.424604455925088,
      "grad_norm": 0.7609820365905762,
      "learning_rate": 0.0002749488201702403,
      "loss": 3.3446,
      "step": 50400
    },
    {
      "epoch": 5.429986007964697,
      "grad_norm": 0.721838116645813,
      "learning_rate": 0.0002746255791401788,
      "loss": 3.3271,
      "step": 50450
    },
    {
      "epoch": 5.435367560004305,
      "grad_norm": 0.7337953448295593,
      "learning_rate": 0.0002743023381101174,
      "loss": 3.3351,
      "step": 50500
    },
    {
      "epoch": 5.440749112043913,
      "grad_norm": 0.7078254818916321,
      "learning_rate": 0.000273979097080056,
      "loss": 3.326,
      "step": 50550
    },
    {
      "epoch": 5.446130664083522,
      "grad_norm": 0.6884511709213257,
      "learning_rate": 0.0002736558560499946,
      "loss": 3.3364,
      "step": 50600
    },
    {
      "epoch": 5.45151221612313,
      "grad_norm": 0.6717587113380432,
      "learning_rate": 0.0002733326150199332,
      "loss": 3.3209,
      "step": 50650
    },
    {
      "epoch": 5.456893768162738,
      "grad_norm": 0.7340549826622009,
      "learning_rate": 0.00027300937398987173,
      "loss": 3.3344,
      "step": 50700
    },
    {
      "epoch": 5.462275320202346,
      "grad_norm": 0.7046282291412354,
      "learning_rate": 0.0002726861329598104,
      "loss": 3.3286,
      "step": 50750
    },
    {
      "epoch": 5.467656872241955,
      "grad_norm": 0.7176598906517029,
      "learning_rate": 0.0002723628919297489,
      "loss": 3.335,
      "step": 50800
    },
    {
      "epoch": 5.473038424281563,
      "grad_norm": 0.7335805296897888,
      "learning_rate": 0.0002720396508996875,
      "loss": 3.3395,
      "step": 50850
    },
    {
      "epoch": 5.478419976321171,
      "grad_norm": 0.7153794169425964,
      "learning_rate": 0.0002717164098696261,
      "loss": 3.3335,
      "step": 50900
    },
    {
      "epoch": 5.483801528360779,
      "grad_norm": 0.7129613757133484,
      "learning_rate": 0.00027139316883956465,
      "loss": 3.34,
      "step": 50950
    },
    {
      "epoch": 5.489183080400387,
      "grad_norm": 0.701945424079895,
      "learning_rate": 0.00027106992780950325,
      "loss": 3.3306,
      "step": 51000
    },
    {
      "epoch": 5.489183080400387,
      "eval_accuracy": 0.3824516459337037,
      "eval_loss": 3.3981103897094727,
      "eval_runtime": 184.7536,
      "eval_samples_per_second": 97.487,
      "eval_steps_per_second": 6.095,
      "step": 51000
    },
    {
      "epoch": 5.494564632439996,
      "grad_norm": 0.6667730212211609,
      "learning_rate": 0.00027074668677944184,
      "loss": 3.3239,
      "step": 51050
    },
    {
      "epoch": 5.499946184479604,
      "grad_norm": 0.6833287477493286,
      "learning_rate": 0.00027042344574938044,
      "loss": 3.33,
      "step": 51100
    },
    {
      "epoch": 5.505327736519212,
      "grad_norm": 0.7144169211387634,
      "learning_rate": 0.000270100204719319,
      "loss": 3.3382,
      "step": 51150
    },
    {
      "epoch": 5.510709288558821,
      "grad_norm": 0.7473929524421692,
      "learning_rate": 0.0002697769636892576,
      "loss": 3.3129,
      "step": 51200
    },
    {
      "epoch": 5.516090840598428,
      "grad_norm": 0.7221124172210693,
      "learning_rate": 0.00026945372265919617,
      "loss": 3.3577,
      "step": 51250
    },
    {
      "epoch": 5.521472392638037,
      "grad_norm": 0.7008888721466064,
      "learning_rate": 0.00026913048162913476,
      "loss": 3.315,
      "step": 51300
    },
    {
      "epoch": 5.5268539446776455,
      "grad_norm": 0.7081114649772644,
      "learning_rate": 0.00026880724059907336,
      "loss": 3.3402,
      "step": 51350
    },
    {
      "epoch": 5.532235496717253,
      "grad_norm": 0.7221799492835999,
      "learning_rate": 0.00026848399956901195,
      "loss": 3.3182,
      "step": 51400
    },
    {
      "epoch": 5.537617048756862,
      "grad_norm": 0.6946343779563904,
      "learning_rate": 0.0002681607585389505,
      "loss": 3.3361,
      "step": 51450
    },
    {
      "epoch": 5.542998600796469,
      "grad_norm": 0.702681303024292,
      "learning_rate": 0.0002678375175088891,
      "loss": 3.3399,
      "step": 51500
    },
    {
      "epoch": 5.548380152836078,
      "grad_norm": 0.7525026798248291,
      "learning_rate": 0.0002675142764788277,
      "loss": 3.3242,
      "step": 51550
    },
    {
      "epoch": 5.553761704875686,
      "grad_norm": 0.772919774055481,
      "learning_rate": 0.0002671910354487663,
      "loss": 3.322,
      "step": 51600
    },
    {
      "epoch": 5.559143256915294,
      "grad_norm": 0.6999807953834534,
      "learning_rate": 0.00026686779441870487,
      "loss": 3.3294,
      "step": 51650
    },
    {
      "epoch": 5.564524808954903,
      "grad_norm": 0.7455286979675293,
      "learning_rate": 0.0002665445533886434,
      "loss": 3.3344,
      "step": 51700
    },
    {
      "epoch": 5.569906360994511,
      "grad_norm": 0.7175178527832031,
      "learning_rate": 0.00026622131235858206,
      "loss": 3.3322,
      "step": 51750
    },
    {
      "epoch": 5.575287913034119,
      "grad_norm": 0.7250377535820007,
      "learning_rate": 0.0002658980713285206,
      "loss": 3.3416,
      "step": 51800
    },
    {
      "epoch": 5.580669465073727,
      "grad_norm": 0.7095009088516235,
      "learning_rate": 0.0002655748302984592,
      "loss": 3.3336,
      "step": 51850
    },
    {
      "epoch": 5.586051017113336,
      "grad_norm": 0.757713258266449,
      "learning_rate": 0.0002652515892683978,
      "loss": 3.3384,
      "step": 51900
    },
    {
      "epoch": 5.591432569152944,
      "grad_norm": 0.7264611124992371,
      "learning_rate": 0.0002649283482383364,
      "loss": 3.3279,
      "step": 51950
    },
    {
      "epoch": 5.596814121192552,
      "grad_norm": 0.7121990323066711,
      "learning_rate": 0.0002646051072082749,
      "loss": 3.3177,
      "step": 52000
    },
    {
      "epoch": 5.596814121192552,
      "eval_accuracy": 0.3827728238700833,
      "eval_loss": 3.3929855823516846,
      "eval_runtime": 184.7294,
      "eval_samples_per_second": 97.499,
      "eval_steps_per_second": 6.095,
      "step": 52000
    },
    {
      "epoch": 5.60219567323216,
      "grad_norm": 0.7471087574958801,
      "learning_rate": 0.0002642818661782135,
      "loss": 3.3679,
      "step": 52050
    },
    {
      "epoch": 5.607577225271768,
      "grad_norm": 0.7511977553367615,
      "learning_rate": 0.0002639586251481521,
      "loss": 3.3404,
      "step": 52100
    },
    {
      "epoch": 5.612958777311377,
      "grad_norm": 0.6835429072380066,
      "learning_rate": 0.0002636353841180907,
      "loss": 3.3509,
      "step": 52150
    },
    {
      "epoch": 5.618340329350985,
      "grad_norm": 0.7123720645904541,
      "learning_rate": 0.0002633121430880293,
      "loss": 3.3175,
      "step": 52200
    },
    {
      "epoch": 5.623721881390593,
      "grad_norm": 0.7043327689170837,
      "learning_rate": 0.00026298890205796784,
      "loss": 3.3321,
      "step": 52250
    },
    {
      "epoch": 5.629103433430201,
      "grad_norm": 0.7442822456359863,
      "learning_rate": 0.00026266566102790644,
      "loss": 3.3466,
      "step": 52300
    },
    {
      "epoch": 5.634484985469809,
      "grad_norm": 0.6664136648178101,
      "learning_rate": 0.00026234241999784503,
      "loss": 3.3312,
      "step": 52350
    },
    {
      "epoch": 5.639866537509418,
      "grad_norm": 0.7649520039558411,
      "learning_rate": 0.0002620191789677836,
      "loss": 3.3195,
      "step": 52400
    },
    {
      "epoch": 5.645248089549026,
      "grad_norm": 0.717771053314209,
      "learning_rate": 0.00026169593793772217,
      "loss": 3.3188,
      "step": 52450
    },
    {
      "epoch": 5.650629641588634,
      "grad_norm": 0.7109053134918213,
      "learning_rate": 0.0002613726969076608,
      "loss": 3.3444,
      "step": 52500
    },
    {
      "epoch": 5.656011193628243,
      "grad_norm": 0.6929636001586914,
      "learning_rate": 0.00026104945587759936,
      "loss": 3.3419,
      "step": 52550
    },
    {
      "epoch": 5.66139274566785,
      "grad_norm": 0.7861093282699585,
      "learning_rate": 0.00026072621484753795,
      "loss": 3.3373,
      "step": 52600
    },
    {
      "epoch": 5.666774297707459,
      "grad_norm": 0.7142403721809387,
      "learning_rate": 0.00026040297381747655,
      "loss": 3.3326,
      "step": 52650
    },
    {
      "epoch": 5.672155849747067,
      "grad_norm": 0.7610068321228027,
      "learning_rate": 0.00026007973278741514,
      "loss": 3.3319,
      "step": 52700
    },
    {
      "epoch": 5.677537401786675,
      "grad_norm": 0.7404228448867798,
      "learning_rate": 0.00025975649175735373,
      "loss": 3.3378,
      "step": 52750
    },
    {
      "epoch": 5.682918953826284,
      "grad_norm": 0.6821990013122559,
      "learning_rate": 0.0002594332507272923,
      "loss": 3.3357,
      "step": 52800
    },
    {
      "epoch": 5.688300505865891,
      "grad_norm": 0.7214856743812561,
      "learning_rate": 0.00025911000969723087,
      "loss": 3.3244,
      "step": 52850
    },
    {
      "epoch": 5.6936820579055,
      "grad_norm": 0.7556411623954773,
      "learning_rate": 0.00025878676866716946,
      "loss": 3.3249,
      "step": 52900
    },
    {
      "epoch": 5.699063609945108,
      "grad_norm": 0.7074944376945496,
      "learning_rate": 0.00025846352763710806,
      "loss": 3.3231,
      "step": 52950
    },
    {
      "epoch": 5.704445161984716,
      "grad_norm": 0.7696301341056824,
      "learning_rate": 0.00025814675142764787,
      "loss": 3.3337,
      "step": 53000
    },
    {
      "epoch": 5.704445161984716,
      "eval_accuracy": 0.38316332234888856,
      "eval_loss": 3.3885035514831543,
      "eval_runtime": 184.6186,
      "eval_samples_per_second": 97.558,
      "eval_steps_per_second": 6.099,
      "step": 53000
    },
    {
      "epoch": 5.709826714024325,
      "grad_norm": 0.7107107043266296,
      "learning_rate": 0.00025782351039758646,
      "loss": 3.3093,
      "step": 53050
    },
    {
      "epoch": 5.715208266063933,
      "grad_norm": 0.7684431672096252,
      "learning_rate": 0.000257500269367525,
      "loss": 3.3285,
      "step": 53100
    },
    {
      "epoch": 5.720589818103541,
      "grad_norm": 0.7184401154518127,
      "learning_rate": 0.0002571770283374636,
      "loss": 3.3186,
      "step": 53150
    },
    {
      "epoch": 5.725971370143149,
      "grad_norm": 0.6955456733703613,
      "learning_rate": 0.0002568537873074022,
      "loss": 3.3399,
      "step": 53200
    },
    {
      "epoch": 5.731352922182758,
      "grad_norm": 0.7292144298553467,
      "learning_rate": 0.0002565305462773408,
      "loss": 3.3194,
      "step": 53250
    },
    {
      "epoch": 5.736734474222366,
      "grad_norm": 0.7374510169029236,
      "learning_rate": 0.0002562073052472794,
      "loss": 3.3199,
      "step": 53300
    },
    {
      "epoch": 5.742116026261974,
      "grad_norm": 0.6997143030166626,
      "learning_rate": 0.0002558840642172179,
      "loss": 3.3359,
      "step": 53350
    },
    {
      "epoch": 5.747497578301582,
      "grad_norm": 0.755856454372406,
      "learning_rate": 0.00025556082318715657,
      "loss": 3.3198,
      "step": 53400
    },
    {
      "epoch": 5.75287913034119,
      "grad_norm": 0.7310744524002075,
      "learning_rate": 0.0002552375821570951,
      "loss": 3.3347,
      "step": 53450
    },
    {
      "epoch": 5.758260682380799,
      "grad_norm": 0.767074704170227,
      "learning_rate": 0.0002549143411270337,
      "loss": 3.3416,
      "step": 53500
    },
    {
      "epoch": 5.763642234420407,
      "grad_norm": 0.7051040530204773,
      "learning_rate": 0.0002545911000969723,
      "loss": 3.3338,
      "step": 53550
    },
    {
      "epoch": 5.769023786460015,
      "grad_norm": 0.7256135940551758,
      "learning_rate": 0.0002542678590669109,
      "loss": 3.3486,
      "step": 53600
    },
    {
      "epoch": 5.774405338499624,
      "grad_norm": 0.7255988121032715,
      "learning_rate": 0.00025394461803684943,
      "loss": 3.3406,
      "step": 53650
    },
    {
      "epoch": 5.779786890539231,
      "grad_norm": 0.7210702300071716,
      "learning_rate": 0.00025362137700678803,
      "loss": 3.3282,
      "step": 53700
    },
    {
      "epoch": 5.78516844257884,
      "grad_norm": 0.7039930820465088,
      "learning_rate": 0.0002532981359767266,
      "loss": 3.3353,
      "step": 53750
    },
    {
      "epoch": 5.790549994618448,
      "grad_norm": 0.7262842059135437,
      "learning_rate": 0.0002529748949466652,
      "loss": 3.3549,
      "step": 53800
    },
    {
      "epoch": 5.795931546658056,
      "grad_norm": 0.718593180179596,
      "learning_rate": 0.0002526516539166038,
      "loss": 3.3341,
      "step": 53850
    },
    {
      "epoch": 5.801313098697665,
      "grad_norm": 0.7095365524291992,
      "learning_rate": 0.00025232841288654235,
      "loss": 3.3327,
      "step": 53900
    },
    {
      "epoch": 5.806694650737272,
      "grad_norm": 0.7695627212524414,
      "learning_rate": 0.00025200517185648095,
      "loss": 3.3315,
      "step": 53950
    },
    {
      "epoch": 5.812076202776881,
      "grad_norm": 0.699380099773407,
      "learning_rate": 0.00025168193082641954,
      "loss": 3.344,
      "step": 54000
    },
    {
      "epoch": 5.812076202776881,
      "eval_accuracy": 0.38343777954350794,
      "eval_loss": 3.3852038383483887,
      "eval_runtime": 184.9031,
      "eval_samples_per_second": 97.408,
      "eval_steps_per_second": 6.09,
      "step": 54000
    },
    {
      "epoch": 5.817457754816489,
      "grad_norm": 0.7718612551689148,
      "learning_rate": 0.00025135868979635814,
      "loss": 3.3475,
      "step": 54050
    },
    {
      "epoch": 5.822839306856097,
      "grad_norm": 0.7421673536300659,
      "learning_rate": 0.00025104191358689794,
      "loss": 3.3234,
      "step": 54100
    },
    {
      "epoch": 5.828220858895706,
      "grad_norm": 0.728031575679779,
      "learning_rate": 0.00025071867255683654,
      "loss": 3.3411,
      "step": 54150
    },
    {
      "epoch": 5.833602410935313,
      "grad_norm": 0.7653467655181885,
      "learning_rate": 0.00025039543152677513,
      "loss": 3.3354,
      "step": 54200
    },
    {
      "epoch": 5.838983962974922,
      "grad_norm": 0.6954116821289062,
      "learning_rate": 0.0002500721904967137,
      "loss": 3.3503,
      "step": 54250
    },
    {
      "epoch": 5.84436551501453,
      "grad_norm": 0.7203173637390137,
      "learning_rate": 0.00024974894946665227,
      "loss": 3.3408,
      "step": 54300
    },
    {
      "epoch": 5.849747067054138,
      "grad_norm": 0.7083039879798889,
      "learning_rate": 0.00024942570843659086,
      "loss": 3.3285,
      "step": 54350
    },
    {
      "epoch": 5.855128619093747,
      "grad_norm": 0.8395054936408997,
      "learning_rate": 0.00024910246740652946,
      "loss": 3.3423,
      "step": 54400
    },
    {
      "epoch": 5.860510171133355,
      "grad_norm": 0.7595872282981873,
      "learning_rate": 0.000248779226376468,
      "loss": 3.3395,
      "step": 54450
    },
    {
      "epoch": 5.865891723172963,
      "grad_norm": 0.7334976196289062,
      "learning_rate": 0.00024845598534640665,
      "loss": 3.3299,
      "step": 54500
    },
    {
      "epoch": 5.871273275212571,
      "grad_norm": 0.7502398490905762,
      "learning_rate": 0.0002481327443163452,
      "loss": 3.3213,
      "step": 54550
    },
    {
      "epoch": 5.87665482725218,
      "grad_norm": 0.8335617184638977,
      "learning_rate": 0.0002478095032862838,
      "loss": 3.3317,
      "step": 54600
    },
    {
      "epoch": 5.882036379291788,
      "grad_norm": 0.7220059037208557,
      "learning_rate": 0.0002474862622562224,
      "loss": 3.3367,
      "step": 54650
    },
    {
      "epoch": 5.887417931331396,
      "grad_norm": 0.7250782251358032,
      "learning_rate": 0.00024716302122616097,
      "loss": 3.3122,
      "step": 54700
    },
    {
      "epoch": 5.892799483371004,
      "grad_norm": 0.7305442690849304,
      "learning_rate": 0.0002468397801960995,
      "loss": 3.3405,
      "step": 54750
    },
    {
      "epoch": 5.898181035410612,
      "grad_norm": 0.6805377006530762,
      "learning_rate": 0.0002465165391660381,
      "loss": 3.3371,
      "step": 54800
    },
    {
      "epoch": 5.903562587450221,
      "grad_norm": 0.7381418347358704,
      "learning_rate": 0.0002461932981359767,
      "loss": 3.3307,
      "step": 54850
    },
    {
      "epoch": 5.9089441394898286,
      "grad_norm": 0.7628136873245239,
      "learning_rate": 0.0002458700571059153,
      "loss": 3.3325,
      "step": 54900
    },
    {
      "epoch": 5.914325691529437,
      "grad_norm": 0.6967856884002686,
      "learning_rate": 0.0002455468160758539,
      "loss": 3.3264,
      "step": 54950
    },
    {
      "epoch": 5.919707243569046,
      "grad_norm": 0.7036657333374023,
      "learning_rate": 0.00024522357504579243,
      "loss": 3.3345,
      "step": 55000
    },
    {
      "epoch": 5.919707243569046,
      "eval_accuracy": 0.3839816958999072,
      "eval_loss": 3.3790087699890137,
      "eval_runtime": 184.9878,
      "eval_samples_per_second": 97.363,
      "eval_steps_per_second": 6.087,
      "step": 55000
    },
    {
      "epoch": 5.925088795608653,
      "grad_norm": 0.7337253093719482,
      "learning_rate": 0.000244900334015731,
      "loss": 3.333,
      "step": 55050
    },
    {
      "epoch": 5.930470347648262,
      "grad_norm": 0.7193291783332825,
      "learning_rate": 0.0002445770929856696,
      "loss": 3.3383,
      "step": 55100
    },
    {
      "epoch": 5.93585189968787,
      "grad_norm": 0.6923545598983765,
      "learning_rate": 0.0002442538519556082,
      "loss": 3.3076,
      "step": 55150
    },
    {
      "epoch": 5.941233451727478,
      "grad_norm": 0.7571631073951721,
      "learning_rate": 0.0002439306109255468,
      "loss": 3.3281,
      "step": 55200
    },
    {
      "epoch": 5.946615003767087,
      "grad_norm": 0.7402598857879639,
      "learning_rate": 0.00024360736989548538,
      "loss": 3.333,
      "step": 55250
    },
    {
      "epoch": 5.951996555806694,
      "grad_norm": 0.7690083384513855,
      "learning_rate": 0.00024328412886542394,
      "loss": 3.3345,
      "step": 55300
    },
    {
      "epoch": 5.957378107846303,
      "grad_norm": 0.6994602084159851,
      "learning_rate": 0.00024296088783536257,
      "loss": 3.3188,
      "step": 55350
    },
    {
      "epoch": 5.962759659885911,
      "grad_norm": 0.7294467091560364,
      "learning_rate": 0.00024263764680530113,
      "loss": 3.3436,
      "step": 55400
    },
    {
      "epoch": 5.968141211925519,
      "grad_norm": 0.7167765498161316,
      "learning_rate": 0.0002423144057752397,
      "loss": 3.3152,
      "step": 55450
    },
    {
      "epoch": 5.973522763965128,
      "grad_norm": 0.733696699142456,
      "learning_rate": 0.00024199116474517832,
      "loss": 3.3195,
      "step": 55500
    },
    {
      "epoch": 5.978904316004736,
      "grad_norm": 0.7189485430717468,
      "learning_rate": 0.0002416679237151169,
      "loss": 3.3247,
      "step": 55550
    },
    {
      "epoch": 5.984285868044344,
      "grad_norm": 0.7302901148796082,
      "learning_rate": 0.00024134468268505546,
      "loss": 3.3269,
      "step": 55600
    },
    {
      "epoch": 5.989667420083952,
      "grad_norm": 0.7239100337028503,
      "learning_rate": 0.00024102144165499405,
      "loss": 3.3427,
      "step": 55650
    },
    {
      "epoch": 5.995048972123561,
      "grad_norm": 0.7289806604385376,
      "learning_rate": 0.00024069820062493265,
      "loss": 3.3162,
      "step": 55700
    },
    {
      "epoch": 6.000430524163169,
      "grad_norm": 0.7002869248390198,
      "learning_rate": 0.00024037495959487121,
      "loss": 3.3247,
      "step": 55750
    },
    {
      "epoch": 6.005812076202777,
      "grad_norm": 0.7519763708114624,
      "learning_rate": 0.0002400517185648098,
      "loss": 3.2449,
      "step": 55800
    },
    {
      "epoch": 6.011193628242385,
      "grad_norm": 0.7565599679946899,
      "learning_rate": 0.00023972847753474838,
      "loss": 3.2444,
      "step": 55850
    },
    {
      "epoch": 6.016575180281993,
      "grad_norm": 0.7228876352310181,
      "learning_rate": 0.00023940523650468697,
      "loss": 3.2454,
      "step": 55900
    },
    {
      "epoch": 6.021956732321602,
      "grad_norm": 0.7430625557899475,
      "learning_rate": 0.00023908199547462557,
      "loss": 3.26,
      "step": 55950
    },
    {
      "epoch": 6.0273382843612096,
      "grad_norm": 0.7767698764801025,
      "learning_rate": 0.00023875875444456413,
      "loss": 3.2586,
      "step": 56000
    },
    {
      "epoch": 6.0273382843612096,
      "eval_accuracy": 0.3844594426476275,
      "eval_loss": 3.383514642715454,
      "eval_runtime": 184.7551,
      "eval_samples_per_second": 97.486,
      "eval_steps_per_second": 6.095,
      "step": 56000
    },
    {
      "epoch": 6.032719836400818,
      "grad_norm": 0.7343913316726685,
      "learning_rate": 0.0002384355134145027,
      "loss": 3.2529,
      "step": 56050
    },
    {
      "epoch": 6.038101388440427,
      "grad_norm": 0.745909571647644,
      "learning_rate": 0.00023811227238444132,
      "loss": 3.242,
      "step": 56100
    },
    {
      "epoch": 6.043482940480034,
      "grad_norm": 0.7055720090866089,
      "learning_rate": 0.0002377890313543799,
      "loss": 3.2403,
      "step": 56150
    },
    {
      "epoch": 6.048864492519643,
      "grad_norm": 0.7497826218605042,
      "learning_rate": 0.00023746579032431849,
      "loss": 3.243,
      "step": 56200
    },
    {
      "epoch": 6.0542460445592505,
      "grad_norm": 0.7048562169075012,
      "learning_rate": 0.00023714254929425708,
      "loss": 3.242,
      "step": 56250
    },
    {
      "epoch": 6.059627596598859,
      "grad_norm": 0.7660420536994934,
      "learning_rate": 0.00023681930826419565,
      "loss": 3.2648,
      "step": 56300
    },
    {
      "epoch": 6.065009148638468,
      "grad_norm": 0.7854742407798767,
      "learning_rate": 0.00023649606723413424,
      "loss": 3.2551,
      "step": 56350
    },
    {
      "epoch": 6.070390700678075,
      "grad_norm": 0.7162970900535583,
      "learning_rate": 0.0002361728262040728,
      "loss": 3.2563,
      "step": 56400
    },
    {
      "epoch": 6.075772252717684,
      "grad_norm": 0.7376888990402222,
      "learning_rate": 0.00023584958517401138,
      "loss": 3.2507,
      "step": 56450
    },
    {
      "epoch": 6.081153804757292,
      "grad_norm": 0.7727888226509094,
      "learning_rate": 0.00023552634414395,
      "loss": 3.2685,
      "step": 56500
    },
    {
      "epoch": 6.0865353567969,
      "grad_norm": 0.7515780925750732,
      "learning_rate": 0.00023520310311388857,
      "loss": 3.2594,
      "step": 56550
    },
    {
      "epoch": 6.091916908836509,
      "grad_norm": 0.721759557723999,
      "learning_rate": 0.00023487986208382713,
      "loss": 3.2608,
      "step": 56600
    },
    {
      "epoch": 6.097298460876116,
      "grad_norm": 0.6812422871589661,
      "learning_rate": 0.00023455662105376576,
      "loss": 3.2503,
      "step": 56650
    },
    {
      "epoch": 6.102680012915725,
      "grad_norm": 0.7772680521011353,
      "learning_rate": 0.00023423338002370432,
      "loss": 3.2606,
      "step": 56700
    },
    {
      "epoch": 6.108061564955333,
      "grad_norm": 0.7657935619354248,
      "learning_rate": 0.0002339101389936429,
      "loss": 3.2543,
      "step": 56750
    },
    {
      "epoch": 6.113443116994941,
      "grad_norm": 0.7089388370513916,
      "learning_rate": 0.00023358689796358149,
      "loss": 3.2707,
      "step": 56800
    },
    {
      "epoch": 6.11882466903455,
      "grad_norm": 0.7366516590118408,
      "learning_rate": 0.00023326365693352008,
      "loss": 3.2698,
      "step": 56850
    },
    {
      "epoch": 6.124206221074158,
      "grad_norm": 0.7500792145729065,
      "learning_rate": 0.00023294041590345865,
      "loss": 3.2517,
      "step": 56900
    },
    {
      "epoch": 6.129587773113766,
      "grad_norm": 0.7379442453384399,
      "learning_rate": 0.00023261717487339724,
      "loss": 3.2377,
      "step": 56950
    },
    {
      "epoch": 6.134969325153374,
      "grad_norm": 0.7361008524894714,
      "learning_rate": 0.0002322939338433358,
      "loss": 3.2657,
      "step": 57000
    },
    {
      "epoch": 6.134969325153374,
      "eval_accuracy": 0.3848917724882413,
      "eval_loss": 3.3805084228515625,
      "eval_runtime": 184.8323,
      "eval_samples_per_second": 97.445,
      "eval_steps_per_second": 6.092,
      "step": 57000
    },
    {
      "epoch": 6.140350877192983,
      "grad_norm": 0.7512038350105286,
      "learning_rate": 0.0002319706928132744,
      "loss": 3.2544,
      "step": 57050
    },
    {
      "epoch": 6.1457324292325906,
      "grad_norm": 0.7466521859169006,
      "learning_rate": 0.000231647451783213,
      "loss": 3.2559,
      "step": 57100
    },
    {
      "epoch": 6.151113981272199,
      "grad_norm": 0.7655845880508423,
      "learning_rate": 0.00023132421075315157,
      "loss": 3.2716,
      "step": 57150
    },
    {
      "epoch": 6.156495533311807,
      "grad_norm": 0.7715071439743042,
      "learning_rate": 0.0002310009697230902,
      "loss": 3.2597,
      "step": 57200
    },
    {
      "epoch": 6.161877085351415,
      "grad_norm": 0.7462551593780518,
      "learning_rate": 0.00023067772869302876,
      "loss": 3.2619,
      "step": 57250
    },
    {
      "epoch": 6.167258637391024,
      "grad_norm": 0.6912649273872375,
      "learning_rate": 0.00023035448766296732,
      "loss": 3.254,
      "step": 57300
    },
    {
      "epoch": 6.1726401894306315,
      "grad_norm": 0.7506750822067261,
      "learning_rate": 0.00023003124663290592,
      "loss": 3.2706,
      "step": 57350
    },
    {
      "epoch": 6.17802174147024,
      "grad_norm": 0.7650833129882812,
      "learning_rate": 0.0002297080056028445,
      "loss": 3.2581,
      "step": 57400
    },
    {
      "epoch": 6.183403293509849,
      "grad_norm": 0.7462422251701355,
      "learning_rate": 0.00022938476457278308,
      "loss": 3.2885,
      "step": 57450
    },
    {
      "epoch": 6.188784845549456,
      "grad_norm": 0.7575925588607788,
      "learning_rate": 0.00022906152354272168,
      "loss": 3.2697,
      "step": 57500
    },
    {
      "epoch": 6.194166397589065,
      "grad_norm": 0.7644687294960022,
      "learning_rate": 0.00022873828251266024,
      "loss": 3.2552,
      "step": 57550
    },
    {
      "epoch": 6.1995479496286725,
      "grad_norm": 0.7716279625892639,
      "learning_rate": 0.00022841504148259884,
      "loss": 3.2557,
      "step": 57600
    },
    {
      "epoch": 6.204929501668281,
      "grad_norm": 0.7543048858642578,
      "learning_rate": 0.00022809180045253743,
      "loss": 3.2624,
      "step": 57650
    },
    {
      "epoch": 6.21031105370789,
      "grad_norm": 0.7387022376060486,
      "learning_rate": 0.000227768559422476,
      "loss": 3.2646,
      "step": 57700
    },
    {
      "epoch": 6.215692605747497,
      "grad_norm": 0.8519375324249268,
      "learning_rate": 0.00022744531839241457,
      "loss": 3.2812,
      "step": 57750
    },
    {
      "epoch": 6.221074157787106,
      "grad_norm": 0.7521229982376099,
      "learning_rate": 0.0002271220773623532,
      "loss": 3.2558,
      "step": 57800
    },
    {
      "epoch": 6.226455709826714,
      "grad_norm": 0.7615373134613037,
      "learning_rate": 0.00022679883633229176,
      "loss": 3.2676,
      "step": 57850
    },
    {
      "epoch": 6.231837261866322,
      "grad_norm": 0.7787977457046509,
      "learning_rate": 0.00022647559530223032,
      "loss": 3.2631,
      "step": 57900
    },
    {
      "epoch": 6.237218813905931,
      "grad_norm": 0.7316073179244995,
      "learning_rate": 0.00022615235427216895,
      "loss": 3.2677,
      "step": 57950
    },
    {
      "epoch": 6.242600365945538,
      "grad_norm": 0.7252430319786072,
      "learning_rate": 0.0002258291132421075,
      "loss": 3.2792,
      "step": 58000
    },
    {
      "epoch": 6.242600365945538,
      "eval_accuracy": 0.38463469975567227,
      "eval_loss": 3.379333019256592,
      "eval_runtime": 184.8261,
      "eval_samples_per_second": 97.448,
      "eval_steps_per_second": 6.092,
      "step": 58000
    },
    {
      "epoch": 6.247981917985147,
      "grad_norm": 0.7599756121635437,
      "learning_rate": 0.0002255058722120461,
      "loss": 3.2763,
      "step": 58050
    },
    {
      "epoch": 6.253363470024755,
      "grad_norm": 0.717863142490387,
      "learning_rate": 0.0002251890960025859,
      "loss": 3.2631,
      "step": 58100
    },
    {
      "epoch": 6.258745022064363,
      "grad_norm": 0.7755172848701477,
      "learning_rate": 0.0002248658549725245,
      "loss": 3.2734,
      "step": 58150
    },
    {
      "epoch": 6.264126574103972,
      "grad_norm": 0.737862229347229,
      "learning_rate": 0.00022454261394246308,
      "loss": 3.2671,
      "step": 58200
    },
    {
      "epoch": 6.26950812614358,
      "grad_norm": 0.7289722561836243,
      "learning_rate": 0.00022421937291240164,
      "loss": 3.2787,
      "step": 58250
    },
    {
      "epoch": 6.274889678183188,
      "grad_norm": 0.737337052822113,
      "learning_rate": 0.00022389613188234027,
      "loss": 3.2799,
      "step": 58300
    },
    {
      "epoch": 6.280271230222796,
      "grad_norm": 0.7309353351593018,
      "learning_rate": 0.00022357289085227883,
      "loss": 3.2639,
      "step": 58350
    },
    {
      "epoch": 6.285652782262405,
      "grad_norm": 0.7572712302207947,
      "learning_rate": 0.0002232496498222174,
      "loss": 3.2858,
      "step": 58400
    },
    {
      "epoch": 6.2910343343020125,
      "grad_norm": 0.7290258407592773,
      "learning_rate": 0.000222926408792156,
      "loss": 3.2798,
      "step": 58450
    },
    {
      "epoch": 6.296415886341621,
      "grad_norm": 0.7298064827919006,
      "learning_rate": 0.0002226031677620946,
      "loss": 3.2858,
      "step": 58500
    },
    {
      "epoch": 6.301797438381229,
      "grad_norm": 0.7554293274879456,
      "learning_rate": 0.00022227992673203316,
      "loss": 3.2744,
      "step": 58550
    },
    {
      "epoch": 6.307178990420837,
      "grad_norm": 0.7588570713996887,
      "learning_rate": 0.00022195668570197175,
      "loss": 3.2767,
      "step": 58600
    },
    {
      "epoch": 6.312560542460446,
      "grad_norm": 0.7721754312515259,
      "learning_rate": 0.00022163344467191032,
      "loss": 3.26,
      "step": 58650
    },
    {
      "epoch": 6.3179420945000535,
      "grad_norm": 0.7513089179992676,
      "learning_rate": 0.00022131020364184891,
      "loss": 3.2749,
      "step": 58700
    },
    {
      "epoch": 6.323323646539662,
      "grad_norm": 0.8285422921180725,
      "learning_rate": 0.0002209869626117875,
      "loss": 3.281,
      "step": 58750
    },
    {
      "epoch": 6.328705198579271,
      "grad_norm": 0.7624416947364807,
      "learning_rate": 0.00022066372158172608,
      "loss": 3.2768,
      "step": 58800
    },
    {
      "epoch": 6.334086750618878,
      "grad_norm": 0.7431687712669373,
      "learning_rate": 0.0002203469453722659,
      "loss": 3.2844,
      "step": 58850
    },
    {
      "epoch": 6.339468302658487,
      "grad_norm": 0.7476500272750854,
      "learning_rate": 0.00022002370434220448,
      "loss": 3.2766,
      "step": 58900
    },
    {
      "epoch": 6.344849854698095,
      "grad_norm": 0.7640836238861084,
      "learning_rate": 0.00021970046331214307,
      "loss": 3.2819,
      "step": 58950
    },
    {
      "epoch": 6.350231406737703,
      "grad_norm": 0.8359128832817078,
      "learning_rate": 0.00021937722228208167,
      "loss": 3.2714,
      "step": 59000
    },
    {
      "epoch": 6.350231406737703,
      "eval_accuracy": 0.38533931373314917,
      "eval_loss": 3.3757359981536865,
      "eval_runtime": 184.9919,
      "eval_samples_per_second": 97.361,
      "eval_steps_per_second": 6.087,
      "step": 59000
    },
    {
      "epoch": 6.355612958777312,
      "grad_norm": 0.7550880312919617,
      "learning_rate": 0.00021905398125202024,
      "loss": 3.2664,
      "step": 59050
    },
    {
      "epoch": 6.360994510816919,
      "grad_norm": 0.7995261549949646,
      "learning_rate": 0.00021873074022195883,
      "loss": 3.2856,
      "step": 59100
    },
    {
      "epoch": 6.366376062856528,
      "grad_norm": 0.7748396992683411,
      "learning_rate": 0.0002184074991918974,
      "loss": 3.2692,
      "step": 59150
    },
    {
      "epoch": 6.371757614896136,
      "grad_norm": 0.7827137112617493,
      "learning_rate": 0.00021808425816183597,
      "loss": 3.298,
      "step": 59200
    },
    {
      "epoch": 6.377139166935744,
      "grad_norm": 0.7864527702331543,
      "learning_rate": 0.0002177610171317746,
      "loss": 3.2718,
      "step": 59250
    },
    {
      "epoch": 6.382520718975353,
      "grad_norm": 0.739658772945404,
      "learning_rate": 0.00021743777610171315,
      "loss": 3.2979,
      "step": 59300
    },
    {
      "epoch": 6.387902271014961,
      "grad_norm": 0.7770798206329346,
      "learning_rate": 0.00021711453507165172,
      "loss": 3.2887,
      "step": 59350
    },
    {
      "epoch": 6.393283823054569,
      "grad_norm": 0.7680968046188354,
      "learning_rate": 0.00021679129404159034,
      "loss": 3.2854,
      "step": 59400
    },
    {
      "epoch": 6.398665375094177,
      "grad_norm": 0.7551103830337524,
      "learning_rate": 0.0002164680530115289,
      "loss": 3.2648,
      "step": 59450
    },
    {
      "epoch": 6.404046927133785,
      "grad_norm": 0.7820691466331482,
      "learning_rate": 0.00021614481198146748,
      "loss": 3.2598,
      "step": 59500
    },
    {
      "epoch": 6.4094284791733935,
      "grad_norm": 0.7943058013916016,
      "learning_rate": 0.00021582157095140607,
      "loss": 3.2879,
      "step": 59550
    },
    {
      "epoch": 6.414810031213002,
      "grad_norm": 0.7810237407684326,
      "learning_rate": 0.00021549832992134467,
      "loss": 3.2768,
      "step": 59600
    },
    {
      "epoch": 6.42019158325261,
      "grad_norm": 0.7318100333213806,
      "learning_rate": 0.00021517508889128324,
      "loss": 3.3002,
      "step": 59650
    },
    {
      "epoch": 6.425573135292218,
      "grad_norm": 0.7224478721618652,
      "learning_rate": 0.00021485184786122183,
      "loss": 3.2621,
      "step": 59700
    },
    {
      "epoch": 6.430954687331827,
      "grad_norm": 0.7693201303482056,
      "learning_rate": 0.0002145286068311604,
      "loss": 3.2572,
      "step": 59750
    },
    {
      "epoch": 6.4363362393714345,
      "grad_norm": 0.7094773650169373,
      "learning_rate": 0.00021420536580109902,
      "loss": 3.2818,
      "step": 59800
    },
    {
      "epoch": 6.441717791411043,
      "grad_norm": 0.761139988899231,
      "learning_rate": 0.0002138821247710376,
      "loss": 3.2627,
      "step": 59850
    },
    {
      "epoch": 6.447099343450651,
      "grad_norm": 0.7221077084541321,
      "learning_rate": 0.00021355888374097615,
      "loss": 3.2619,
      "step": 59900
    },
    {
      "epoch": 6.452480895490259,
      "grad_norm": 0.7425982356071472,
      "learning_rate": 0.00021323564271091478,
      "loss": 3.297,
      "step": 59950
    },
    {
      "epoch": 6.457862447529868,
      "grad_norm": 0.8136957287788391,
      "learning_rate": 0.00021291240168085334,
      "loss": 3.2827,
      "step": 60000
    },
    {
      "epoch": 6.457862447529868,
      "eval_accuracy": 0.3858433544797205,
      "eval_loss": 3.369577169418335,
      "eval_runtime": 184.5983,
      "eval_samples_per_second": 97.569,
      "eval_steps_per_second": 6.1,
      "step": 60000
    },
    {
      "epoch": 6.4632439995694755,
      "grad_norm": 0.7648407220840454,
      "learning_rate": 0.0002125891606507919,
      "loss": 3.2689,
      "step": 60050
    },
    {
      "epoch": 6.468625551609084,
      "grad_norm": 0.7560406923294067,
      "learning_rate": 0.0002122659196207305,
      "loss": 3.2828,
      "step": 60100
    },
    {
      "epoch": 6.474007103648693,
      "grad_norm": 0.7538021206855774,
      "learning_rate": 0.0002119426785906691,
      "loss": 3.2751,
      "step": 60150
    },
    {
      "epoch": 6.4793886556883,
      "grad_norm": 0.796912670135498,
      "learning_rate": 0.00021161943756060767,
      "loss": 3.2776,
      "step": 60200
    },
    {
      "epoch": 6.484770207727909,
      "grad_norm": 0.7616314888000488,
      "learning_rate": 0.00021129619653054626,
      "loss": 3.2964,
      "step": 60250
    },
    {
      "epoch": 6.490151759767517,
      "grad_norm": 0.7025566101074219,
      "learning_rate": 0.00021097295550048483,
      "loss": 3.2876,
      "step": 60300
    },
    {
      "epoch": 6.495533311807125,
      "grad_norm": 0.775004506111145,
      "learning_rate": 0.00021064971447042343,
      "loss": 3.2618,
      "step": 60350
    },
    {
      "epoch": 6.500914863846734,
      "grad_norm": 0.7972927689552307,
      "learning_rate": 0.00021032647344036202,
      "loss": 3.2837,
      "step": 60400
    },
    {
      "epoch": 6.506296415886341,
      "grad_norm": 0.7656959295272827,
      "learning_rate": 0.0002100032324103006,
      "loss": 3.2853,
      "step": 60450
    },
    {
      "epoch": 6.51167796792595,
      "grad_norm": 0.7854245901107788,
      "learning_rate": 0.00020967999138023916,
      "loss": 3.2795,
      "step": 60500
    },
    {
      "epoch": 6.517059519965558,
      "grad_norm": 0.7886183857917786,
      "learning_rate": 0.00020935675035017778,
      "loss": 3.2849,
      "step": 60550
    },
    {
      "epoch": 6.522441072005166,
      "grad_norm": 0.8237834572792053,
      "learning_rate": 0.00020903350932011634,
      "loss": 3.285,
      "step": 60600
    },
    {
      "epoch": 6.5278226240447745,
      "grad_norm": 0.7524319291114807,
      "learning_rate": 0.0002087102682900549,
      "loss": 3.2721,
      "step": 60650
    },
    {
      "epoch": 6.533204176084383,
      "grad_norm": 0.7918885946273804,
      "learning_rate": 0.00020838702725999353,
      "loss": 3.2733,
      "step": 60700
    },
    {
      "epoch": 6.538585728123991,
      "grad_norm": 0.742572546005249,
      "learning_rate": 0.0002080637862299321,
      "loss": 3.2713,
      "step": 60750
    },
    {
      "epoch": 6.543967280163599,
      "grad_norm": 0.7656800150871277,
      "learning_rate": 0.0002077405451998707,
      "loss": 3.2906,
      "step": 60800
    },
    {
      "epoch": 6.549348832203208,
      "grad_norm": 0.7438012361526489,
      "learning_rate": 0.00020741730416980926,
      "loss": 3.2745,
      "step": 60850
    },
    {
      "epoch": 6.5547303842428155,
      "grad_norm": 0.793135404586792,
      "learning_rate": 0.00020709406313974786,
      "loss": 3.2778,
      "step": 60900
    },
    {
      "epoch": 6.560111936282424,
      "grad_norm": 0.7506826519966125,
      "learning_rate": 0.00020677082210968645,
      "loss": 3.2682,
      "step": 60950
    },
    {
      "epoch": 6.565493488322032,
      "grad_norm": 0.7645418047904968,
      "learning_rate": 0.00020644758107962502,
      "loss": 3.2671,
      "step": 61000
    },
    {
      "epoch": 6.565493488322032,
      "eval_accuracy": 0.386022305785951,
      "eval_loss": 3.368492603302002,
      "eval_runtime": 185.0491,
      "eval_samples_per_second": 97.331,
      "eval_steps_per_second": 6.085,
      "step": 61000
    },
    {
      "epoch": 6.57087504036164,
      "grad_norm": 0.7325201630592346,
      "learning_rate": 0.0002061243400495636,
      "loss": 3.2796,
      "step": 61050
    },
    {
      "epoch": 6.576256592401249,
      "grad_norm": 0.7918336987495422,
      "learning_rate": 0.0002058010990195022,
      "loss": 3.2781,
      "step": 61100
    },
    {
      "epoch": 6.5816381444408565,
      "grad_norm": 0.7214813232421875,
      "learning_rate": 0.00020547785798944078,
      "loss": 3.2881,
      "step": 61150
    },
    {
      "epoch": 6.587019696480465,
      "grad_norm": 0.7659515738487244,
      "learning_rate": 0.00020515461695937934,
      "loss": 3.2731,
      "step": 61200
    },
    {
      "epoch": 6.592401248520073,
      "grad_norm": 0.7384321093559265,
      "learning_rate": 0.00020483137592931797,
      "loss": 3.2826,
      "step": 61250
    },
    {
      "epoch": 6.597782800559681,
      "grad_norm": 0.7732715010643005,
      "learning_rate": 0.00020450813489925653,
      "loss": 3.2796,
      "step": 61300
    },
    {
      "epoch": 6.60316435259929,
      "grad_norm": 0.7654476165771484,
      "learning_rate": 0.0002041848938691951,
      "loss": 3.2582,
      "step": 61350
    },
    {
      "epoch": 6.608545904638898,
      "grad_norm": 0.7931333184242249,
      "learning_rate": 0.0002038616528391337,
      "loss": 3.2684,
      "step": 61400
    },
    {
      "epoch": 6.613927456678506,
      "grad_norm": 0.7446388602256775,
      "learning_rate": 0.00020353841180907226,
      "loss": 3.2826,
      "step": 61450
    },
    {
      "epoch": 6.619309008718115,
      "grad_norm": 0.7642561197280884,
      "learning_rate": 0.00020321517077901086,
      "loss": 3.2918,
      "step": 61500
    },
    {
      "epoch": 6.624690560757722,
      "grad_norm": 0.8336501717567444,
      "learning_rate": 0.00020289192974894945,
      "loss": 3.2695,
      "step": 61550
    },
    {
      "epoch": 6.630072112797331,
      "grad_norm": 0.7792569994926453,
      "learning_rate": 0.00020256868871888802,
      "loss": 3.285,
      "step": 61600
    },
    {
      "epoch": 6.635453664836939,
      "grad_norm": 0.8366368412971497,
      "learning_rate": 0.00020224544768882664,
      "loss": 3.2801,
      "step": 61650
    },
    {
      "epoch": 6.640835216876547,
      "grad_norm": 0.7799500226974487,
      "learning_rate": 0.0002019222066587652,
      "loss": 3.2867,
      "step": 61700
    },
    {
      "epoch": 6.6462167689161555,
      "grad_norm": 0.8103165626525879,
      "learning_rate": 0.00020159896562870378,
      "loss": 3.2709,
      "step": 61750
    },
    {
      "epoch": 6.651598320955763,
      "grad_norm": 0.7403290867805481,
      "learning_rate": 0.00020127572459864237,
      "loss": 3.2901,
      "step": 61800
    },
    {
      "epoch": 6.656979872995372,
      "grad_norm": 0.7185531854629517,
      "learning_rate": 0.00020095248356858097,
      "loss": 3.2843,
      "step": 61850
    },
    {
      "epoch": 6.66236142503498,
      "grad_norm": 0.7514901161193848,
      "learning_rate": 0.00020062924253851953,
      "loss": 3.2866,
      "step": 61900
    },
    {
      "epoch": 6.667742977074588,
      "grad_norm": 0.7954921126365662,
      "learning_rate": 0.00020030600150845813,
      "loss": 3.2607,
      "step": 61950
    },
    {
      "epoch": 6.6731245291141965,
      "grad_norm": 0.7863759398460388,
      "learning_rate": 0.0001999827604783967,
      "loss": 3.3058,
      "step": 62000
    },
    {
      "epoch": 6.6731245291141965,
      "eval_accuracy": 0.3863287069295878,
      "eval_loss": 3.3633809089660645,
      "eval_runtime": 184.8165,
      "eval_samples_per_second": 97.453,
      "eval_steps_per_second": 6.093,
      "step": 62000
    },
    {
      "epoch": 6.678506081153805,
      "grad_norm": 0.7923064827919006,
      "learning_rate": 0.0001996595194483353,
      "loss": 3.3203,
      "step": 62050
    },
    {
      "epoch": 6.683887633193413,
      "grad_norm": 0.7636295557022095,
      "learning_rate": 0.00019933627841827389,
      "loss": 3.297,
      "step": 62100
    },
    {
      "epoch": 6.689269185233021,
      "grad_norm": 0.740822970867157,
      "learning_rate": 0.00019901303738821245,
      "loss": 3.2925,
      "step": 62150
    },
    {
      "epoch": 6.69465073727263,
      "grad_norm": 0.7824040651321411,
      "learning_rate": 0.00019868979635815102,
      "loss": 3.2845,
      "step": 62200
    },
    {
      "epoch": 6.7000322893122375,
      "grad_norm": 0.7348474860191345,
      "learning_rate": 0.00019836655532808964,
      "loss": 3.2755,
      "step": 62250
    },
    {
      "epoch": 6.705413841351846,
      "grad_norm": 0.8276709318161011,
      "learning_rate": 0.0001980433142980282,
      "loss": 3.302,
      "step": 62300
    },
    {
      "epoch": 6.710795393391454,
      "grad_norm": 0.7462751865386963,
      "learning_rate": 0.00019772007326796678,
      "loss": 3.2734,
      "step": 62350
    },
    {
      "epoch": 6.716176945431062,
      "grad_norm": 0.8004351258277893,
      "learning_rate": 0.0001973968322379054,
      "loss": 3.278,
      "step": 62400
    },
    {
      "epoch": 6.721558497470671,
      "grad_norm": 0.7743335366249084,
      "learning_rate": 0.00019707359120784397,
      "loss": 3.2841,
      "step": 62450
    },
    {
      "epoch": 6.7269400495102785,
      "grad_norm": 0.7511734366416931,
      "learning_rate": 0.00019675035017778253,
      "loss": 3.277,
      "step": 62500
    },
    {
      "epoch": 6.732321601549887,
      "grad_norm": 0.7555059194564819,
      "learning_rate": 0.00019642710914772113,
      "loss": 3.2914,
      "step": 62550
    },
    {
      "epoch": 6.737703153589496,
      "grad_norm": 0.7607969641685486,
      "learning_rate": 0.00019610386811765972,
      "loss": 3.2923,
      "step": 62600
    },
    {
      "epoch": 6.743084705629103,
      "grad_norm": 0.7745279669761658,
      "learning_rate": 0.00019578062708759832,
      "loss": 3.2707,
      "step": 62650
    },
    {
      "epoch": 6.748466257668712,
      "grad_norm": 0.7459644675254822,
      "learning_rate": 0.00019545738605753689,
      "loss": 3.2656,
      "step": 62700
    },
    {
      "epoch": 6.75384780970832,
      "grad_norm": 0.7880685329437256,
      "learning_rate": 0.00019513414502747545,
      "loss": 3.271,
      "step": 62750
    },
    {
      "epoch": 6.759229361747928,
      "grad_norm": 0.7434306740760803,
      "learning_rate": 0.00019481090399741408,
      "loss": 3.2812,
      "step": 62800
    },
    {
      "epoch": 6.7646109137875365,
      "grad_norm": 0.8286969065666199,
      "learning_rate": 0.0001945005926085551,
      "loss": 3.2948,
      "step": 62850
    },
    {
      "epoch": 6.769992465827144,
      "grad_norm": 0.7807492613792419,
      "learning_rate": 0.0001941773515784937,
      "loss": 3.296,
      "step": 62900
    },
    {
      "epoch": 6.775374017866753,
      "grad_norm": 0.8009621500968933,
      "learning_rate": 0.00019385411054843226,
      "loss": 3.2822,
      "step": 62950
    },
    {
      "epoch": 6.780755569906361,
      "grad_norm": 0.7488981485366821,
      "learning_rate": 0.00019353086951837085,
      "loss": 3.2775,
      "step": 63000
    },
    {
      "epoch": 6.780755569906361,
      "eval_accuracy": 0.3868306832712907,
      "eval_loss": 3.3595683574676514,
      "eval_runtime": 184.9791,
      "eval_samples_per_second": 97.368,
      "eval_steps_per_second": 6.087,
      "step": 63000
    },
    {
      "epoch": 6.786137121945969,
      "grad_norm": 0.7897948026657104,
      "learning_rate": 0.00019320762848830942,
      "loss": 3.277,
      "step": 63050
    },
    {
      "epoch": 6.7915186739855775,
      "grad_norm": 0.7589372992515564,
      "learning_rate": 0.000192884387458248,
      "loss": 3.2875,
      "step": 63100
    },
    {
      "epoch": 6.796900226025185,
      "grad_norm": 0.7196517586708069,
      "learning_rate": 0.0001925611464281866,
      "loss": 3.2844,
      "step": 63150
    },
    {
      "epoch": 6.802281778064794,
      "grad_norm": 0.7708859443664551,
      "learning_rate": 0.00019223790539812518,
      "loss": 3.2796,
      "step": 63200
    },
    {
      "epoch": 6.807663330104402,
      "grad_norm": 0.7655634880065918,
      "learning_rate": 0.00019191466436806374,
      "loss": 3.2928,
      "step": 63250
    },
    {
      "epoch": 6.813044882144011,
      "grad_norm": 0.8020054697990417,
      "learning_rate": 0.00019159142333800236,
      "loss": 3.259,
      "step": 63300
    },
    {
      "epoch": 6.8184264341836185,
      "grad_norm": 0.7741644382476807,
      "learning_rate": 0.00019126818230794093,
      "loss": 3.2918,
      "step": 63350
    },
    {
      "epoch": 6.823807986223227,
      "grad_norm": 0.7251269221305847,
      "learning_rate": 0.00019094494127787953,
      "loss": 3.2822,
      "step": 63400
    },
    {
      "epoch": 6.829189538262835,
      "grad_norm": 0.7983183860778809,
      "learning_rate": 0.00019062170024781812,
      "loss": 3.3032,
      "step": 63450
    },
    {
      "epoch": 6.834571090302443,
      "grad_norm": 0.7949510216712952,
      "learning_rate": 0.0001902984592177567,
      "loss": 3.284,
      "step": 63500
    },
    {
      "epoch": 6.839952642342052,
      "grad_norm": 0.788966715335846,
      "learning_rate": 0.00018997521818769528,
      "loss": 3.2741,
      "step": 63550
    },
    {
      "epoch": 6.8453341943816595,
      "grad_norm": 0.7868644595146179,
      "learning_rate": 0.00018965197715763385,
      "loss": 3.2781,
      "step": 63600
    },
    {
      "epoch": 6.850715746421268,
      "grad_norm": 0.7467278242111206,
      "learning_rate": 0.00018932873612757245,
      "loss": 3.2761,
      "step": 63650
    },
    {
      "epoch": 6.856097298460876,
      "grad_norm": 0.7820574641227722,
      "learning_rate": 0.00018900549509751104,
      "loss": 3.2763,
      "step": 63700
    },
    {
      "epoch": 6.861478850500484,
      "grad_norm": 0.7491395473480225,
      "learning_rate": 0.0001886822540674496,
      "loss": 3.2862,
      "step": 63750
    },
    {
      "epoch": 6.866860402540093,
      "grad_norm": 0.8236872553825378,
      "learning_rate": 0.00018835901303738818,
      "loss": 3.2784,
      "step": 63800
    },
    {
      "epoch": 6.8722419545797,
      "grad_norm": 0.7589198350906372,
      "learning_rate": 0.0001880357720073268,
      "loss": 3.2835,
      "step": 63850
    },
    {
      "epoch": 6.877623506619309,
      "grad_norm": 0.7768234014511108,
      "learning_rate": 0.00018771253097726537,
      "loss": 3.2868,
      "step": 63900
    },
    {
      "epoch": 6.8830050586589175,
      "grad_norm": 0.7730334997177124,
      "learning_rate": 0.00018738928994720393,
      "loss": 3.292,
      "step": 63950
    },
    {
      "epoch": 6.888386610698525,
      "grad_norm": 0.7769656181335449,
      "learning_rate": 0.00018706604891714255,
      "loss": 3.2919,
      "step": 64000
    },
    {
      "epoch": 6.888386610698525,
      "eval_accuracy": 0.3873528690501271,
      "eval_loss": 3.354077100753784,
      "eval_runtime": 185.2307,
      "eval_samples_per_second": 97.236,
      "eval_steps_per_second": 6.079,
      "step": 64000
    },
    {
      "epoch": 6.893768162738134,
      "grad_norm": 0.8213059902191162,
      "learning_rate": 0.00018674280788708112,
      "loss": 3.2788,
      "step": 64050
    },
    {
      "epoch": 6.899149714777742,
      "grad_norm": 0.7441644668579102,
      "learning_rate": 0.0001864195668570197,
      "loss": 3.2803,
      "step": 64100
    },
    {
      "epoch": 6.90453126681735,
      "grad_norm": 0.7904473543167114,
      "learning_rate": 0.00018609632582695828,
      "loss": 3.2874,
      "step": 64150
    },
    {
      "epoch": 6.9099128188569585,
      "grad_norm": 0.7749623656272888,
      "learning_rate": 0.00018577308479689685,
      "loss": 3.2675,
      "step": 64200
    },
    {
      "epoch": 6.915294370896566,
      "grad_norm": 0.7651856541633606,
      "learning_rate": 0.00018544984376683545,
      "loss": 3.2799,
      "step": 64250
    },
    {
      "epoch": 6.920675922936175,
      "grad_norm": 0.8032562136650085,
      "learning_rate": 0.00018512660273677404,
      "loss": 3.2993,
      "step": 64300
    },
    {
      "epoch": 6.926057474975783,
      "grad_norm": 0.7830508351325989,
      "learning_rate": 0.0001848033617067126,
      "loss": 3.2769,
      "step": 64350
    },
    {
      "epoch": 6.931439027015391,
      "grad_norm": 0.778607964515686,
      "learning_rate": 0.00018448012067665123,
      "loss": 3.2785,
      "step": 64400
    },
    {
      "epoch": 6.9368205790549995,
      "grad_norm": 0.7424008846282959,
      "learning_rate": 0.0001841568796465898,
      "loss": 3.2746,
      "step": 64450
    },
    {
      "epoch": 6.942202131094608,
      "grad_norm": 0.7957499027252197,
      "learning_rate": 0.00018383363861652837,
      "loss": 3.2933,
      "step": 64500
    },
    {
      "epoch": 6.947583683134216,
      "grad_norm": 0.8147248029708862,
      "learning_rate": 0.00018351039758646696,
      "loss": 3.273,
      "step": 64550
    },
    {
      "epoch": 6.952965235173824,
      "grad_norm": 0.7838407158851624,
      "learning_rate": 0.00018318715655640555,
      "loss": 3.2845,
      "step": 64600
    },
    {
      "epoch": 6.958346787213433,
      "grad_norm": 0.7947826385498047,
      "learning_rate": 0.00018286391552634412,
      "loss": 3.299,
      "step": 64650
    },
    {
      "epoch": 6.9637283392530405,
      "grad_norm": 0.8972591161727905,
      "learning_rate": 0.00018254067449628272,
      "loss": 3.3021,
      "step": 64700
    },
    {
      "epoch": 6.969109891292649,
      "grad_norm": 0.7764172554016113,
      "learning_rate": 0.00018221743346622128,
      "loss": 3.2696,
      "step": 64750
    },
    {
      "epoch": 6.974491443332257,
      "grad_norm": 0.8290959000587463,
      "learning_rate": 0.00018189419243615988,
      "loss": 3.2747,
      "step": 64800
    },
    {
      "epoch": 6.979872995371865,
      "grad_norm": 0.751587450504303,
      "learning_rate": 0.00018157095140609847,
      "loss": 3.2917,
      "step": 64850
    },
    {
      "epoch": 6.985254547411474,
      "grad_norm": 0.7895004153251648,
      "learning_rate": 0.00018124771037603704,
      "loss": 3.2852,
      "step": 64900
    },
    {
      "epoch": 6.990636099451081,
      "grad_norm": 0.7497310638427734,
      "learning_rate": 0.0001809244693459756,
      "loss": 3.3,
      "step": 64950
    },
    {
      "epoch": 6.99601765149069,
      "grad_norm": 0.763469398021698,
      "learning_rate": 0.00018060122831591423,
      "loss": 3.2603,
      "step": 65000
    },
    {
      "epoch": 6.99601765149069,
      "eval_accuracy": 0.38770675150573886,
      "eval_loss": 3.351874828338623,
      "eval_runtime": 184.7786,
      "eval_samples_per_second": 97.473,
      "eval_steps_per_second": 6.094,
      "step": 65000
    },
    {
      "epoch": 7.0013992035302985,
      "grad_norm": 0.7597163915634155,
      "learning_rate": 0.0001802779872858528,
      "loss": 3.2679,
      "step": 65050
    },
    {
      "epoch": 7.006780755569906,
      "grad_norm": 0.7525449395179749,
      "learning_rate": 0.00017995474625579137,
      "loss": 3.2098,
      "step": 65100
    },
    {
      "epoch": 7.012162307609515,
      "grad_norm": 0.8813150525093079,
      "learning_rate": 0.00017963150522573,
      "loss": 3.1881,
      "step": 65150
    },
    {
      "epoch": 7.017543859649122,
      "grad_norm": 0.787375271320343,
      "learning_rate": 0.0001793147290162698,
      "loss": 3.1957,
      "step": 65200
    },
    {
      "epoch": 7.022925411688731,
      "grad_norm": 0.7959186434745789,
      "learning_rate": 0.00017899148798620836,
      "loss": 3.2153,
      "step": 65250
    },
    {
      "epoch": 7.0283069637283395,
      "grad_norm": 0.7243605256080627,
      "learning_rate": 0.00017866824695614696,
      "loss": 3.1987,
      "step": 65300
    },
    {
      "epoch": 7.033688515767947,
      "grad_norm": 0.7697280049324036,
      "learning_rate": 0.00017834500592608555,
      "loss": 3.2116,
      "step": 65350
    },
    {
      "epoch": 7.039070067807556,
      "grad_norm": 0.781091034412384,
      "learning_rate": 0.00017802176489602412,
      "loss": 3.192,
      "step": 65400
    },
    {
      "epoch": 7.044451619847164,
      "grad_norm": 0.8070541620254517,
      "learning_rate": 0.00017769852386596269,
      "loss": 3.2159,
      "step": 65450
    },
    {
      "epoch": 7.049833171886772,
      "grad_norm": 0.7928311228752136,
      "learning_rate": 0.0001773752828359013,
      "loss": 3.2199,
      "step": 65500
    },
    {
      "epoch": 7.0552147239263805,
      "grad_norm": 0.8367716670036316,
      "learning_rate": 0.00017705204180583988,
      "loss": 3.1962,
      "step": 65550
    },
    {
      "epoch": 7.060596275965988,
      "grad_norm": 0.8304590582847595,
      "learning_rate": 0.00017672880077577844,
      "loss": 3.2124,
      "step": 65600
    },
    {
      "epoch": 7.065977828005597,
      "grad_norm": 0.789740800857544,
      "learning_rate": 0.00017640555974571704,
      "loss": 3.1983,
      "step": 65650
    },
    {
      "epoch": 7.071359380045205,
      "grad_norm": 0.7952477335929871,
      "learning_rate": 0.00017608231871565563,
      "loss": 3.2149,
      "step": 65700
    },
    {
      "epoch": 7.076740932084813,
      "grad_norm": 0.8178505897521973,
      "learning_rate": 0.0001757590776855942,
      "loss": 3.2005,
      "step": 65750
    },
    {
      "epoch": 7.0821224841244215,
      "grad_norm": 0.8267757892608643,
      "learning_rate": 0.0001754358366555328,
      "loss": 3.2145,
      "step": 65800
    },
    {
      "epoch": 7.08750403616403,
      "grad_norm": 0.7804736495018005,
      "learning_rate": 0.00017511259562547136,
      "loss": 3.204,
      "step": 65850
    },
    {
      "epoch": 7.092885588203638,
      "grad_norm": 0.8235890865325928,
      "learning_rate": 0.00017478935459540996,
      "loss": 3.2238,
      "step": 65900
    },
    {
      "epoch": 7.098267140243246,
      "grad_norm": 0.8137307167053223,
      "learning_rate": 0.00017446611356534855,
      "loss": 3.2068,
      "step": 65950
    },
    {
      "epoch": 7.103648692282855,
      "grad_norm": 0.7665497660636902,
      "learning_rate": 0.00017414287253528712,
      "loss": 3.1971,
      "step": 66000
    },
    {
      "epoch": 7.103648692282855,
      "eval_accuracy": 0.3873983946101214,
      "eval_loss": 3.3571224212646484,
      "eval_runtime": 185.0933,
      "eval_samples_per_second": 97.308,
      "eval_steps_per_second": 6.083,
      "step": 66000
    },
    {
      "epoch": 7.109030244322462,
      "grad_norm": 0.8523070812225342,
      "learning_rate": 0.00017381963150522569,
      "loss": 3.2132,
      "step": 66050
    },
    {
      "epoch": 7.114411796362071,
      "grad_norm": 0.8501766920089722,
      "learning_rate": 0.0001734963904751643,
      "loss": 3.2196,
      "step": 66100
    },
    {
      "epoch": 7.119793348401679,
      "grad_norm": 0.7479291558265686,
      "learning_rate": 0.00017317314944510288,
      "loss": 3.2021,
      "step": 66150
    },
    {
      "epoch": 7.125174900441287,
      "grad_norm": 0.8581606149673462,
      "learning_rate": 0.00017284990841504147,
      "loss": 3.238,
      "step": 66200
    },
    {
      "epoch": 7.130556452480896,
      "grad_norm": 0.8038374185562134,
      "learning_rate": 0.00017252666738498007,
      "loss": 3.2067,
      "step": 66250
    },
    {
      "epoch": 7.135938004520503,
      "grad_norm": 0.7955670356750488,
      "learning_rate": 0.00017220342635491863,
      "loss": 3.2062,
      "step": 66300
    },
    {
      "epoch": 7.141319556560112,
      "grad_norm": 0.8133190274238586,
      "learning_rate": 0.00017188018532485723,
      "loss": 3.2189,
      "step": 66350
    },
    {
      "epoch": 7.1467011085997205,
      "grad_norm": 0.7808161377906799,
      "learning_rate": 0.0001715569442947958,
      "loss": 3.2181,
      "step": 66400
    },
    {
      "epoch": 7.152082660639328,
      "grad_norm": 0.7812925577163696,
      "learning_rate": 0.0001712337032647344,
      "loss": 3.21,
      "step": 66450
    },
    {
      "epoch": 7.157464212678937,
      "grad_norm": 0.7803530097007751,
      "learning_rate": 0.00017091046223467298,
      "loss": 3.221,
      "step": 66500
    },
    {
      "epoch": 7.162845764718545,
      "grad_norm": 0.8198121786117554,
      "learning_rate": 0.00017058722120461155,
      "loss": 3.2165,
      "step": 66550
    },
    {
      "epoch": 7.168227316758153,
      "grad_norm": 0.777133047580719,
      "learning_rate": 0.00017026398017455012,
      "loss": 3.2101,
      "step": 66600
    },
    {
      "epoch": 7.1736088687977615,
      "grad_norm": 0.7378299236297607,
      "learning_rate": 0.00016994073914448874,
      "loss": 3.2223,
      "step": 66650
    },
    {
      "epoch": 7.178990420837369,
      "grad_norm": 0.8564599752426147,
      "learning_rate": 0.0001696174981144273,
      "loss": 3.2103,
      "step": 66700
    },
    {
      "epoch": 7.184371972876978,
      "grad_norm": 0.8352597951889038,
      "learning_rate": 0.00016929425708436588,
      "loss": 3.2395,
      "step": 66750
    },
    {
      "epoch": 7.189753524916586,
      "grad_norm": 0.7732496857643127,
      "learning_rate": 0.0001689710160543045,
      "loss": 3.209,
      "step": 66800
    },
    {
      "epoch": 7.195135076956194,
      "grad_norm": 0.7928017377853394,
      "learning_rate": 0.00016864777502424307,
      "loss": 3.208,
      "step": 66850
    },
    {
      "epoch": 7.2005166289958025,
      "grad_norm": 0.8274809122085571,
      "learning_rate": 0.00016832453399418163,
      "loss": 3.2147,
      "step": 66900
    },
    {
      "epoch": 7.205898181035411,
      "grad_norm": 0.812453031539917,
      "learning_rate": 0.00016800129296412023,
      "loss": 3.2238,
      "step": 66950
    },
    {
      "epoch": 7.211279733075019,
      "grad_norm": 0.7714659571647644,
      "learning_rate": 0.00016767805193405882,
      "loss": 3.2284,
      "step": 67000
    },
    {
      "epoch": 7.211279733075019,
      "eval_accuracy": 0.38780638620386476,
      "eval_loss": 3.356039047241211,
      "eval_runtime": 185.0026,
      "eval_samples_per_second": 97.355,
      "eval_steps_per_second": 6.086,
      "step": 67000
    },
    {
      "epoch": 7.216661285114627,
      "grad_norm": 0.7873409390449524,
      "learning_rate": 0.0001673548109039974,
      "loss": 3.2216,
      "step": 67050
    },
    {
      "epoch": 7.222042837154235,
      "grad_norm": 0.8289926052093506,
      "learning_rate": 0.00016703156987393598,
      "loss": 3.2315,
      "step": 67100
    },
    {
      "epoch": 7.2274243891938434,
      "grad_norm": 0.7620259523391724,
      "learning_rate": 0.00016670832884387455,
      "loss": 3.2078,
      "step": 67150
    },
    {
      "epoch": 7.232805941233452,
      "grad_norm": 0.8179631233215332,
      "learning_rate": 0.00016638508781381317,
      "loss": 3.2387,
      "step": 67200
    },
    {
      "epoch": 7.23818749327306,
      "grad_norm": 0.8092691898345947,
      "learning_rate": 0.00016606184678375174,
      "loss": 3.2218,
      "step": 67250
    },
    {
      "epoch": 7.243569045312668,
      "grad_norm": 0.8265263438224792,
      "learning_rate": 0.0001657386057536903,
      "loss": 3.2141,
      "step": 67300
    },
    {
      "epoch": 7.248950597352277,
      "grad_norm": 0.7724214196205139,
      "learning_rate": 0.00016541536472362893,
      "loss": 3.2062,
      "step": 67350
    },
    {
      "epoch": 7.254332149391884,
      "grad_norm": 0.8382948040962219,
      "learning_rate": 0.0001650921236935675,
      "loss": 3.2272,
      "step": 67400
    },
    {
      "epoch": 7.259713701431493,
      "grad_norm": 0.7633907794952393,
      "learning_rate": 0.00016476888266350607,
      "loss": 3.2307,
      "step": 67450
    },
    {
      "epoch": 7.265095253471101,
      "grad_norm": 0.8241746425628662,
      "learning_rate": 0.00016444564163344466,
      "loss": 3.2082,
      "step": 67500
    },
    {
      "epoch": 7.270476805510709,
      "grad_norm": 0.8180881142616272,
      "learning_rate": 0.00016412240060338326,
      "loss": 3.2256,
      "step": 67550
    },
    {
      "epoch": 7.275858357550318,
      "grad_norm": 0.8148139119148254,
      "learning_rate": 0.00016379915957332182,
      "loss": 3.2394,
      "step": 67600
    },
    {
      "epoch": 7.281239909589925,
      "grad_norm": 0.8224004507064819,
      "learning_rate": 0.00016348238336386163,
      "loss": 3.2298,
      "step": 67650
    },
    {
      "epoch": 7.286621461629534,
      "grad_norm": 0.7899991273880005,
      "learning_rate": 0.0001631591423338002,
      "loss": 3.2128,
      "step": 67700
    },
    {
      "epoch": 7.2920030136691425,
      "grad_norm": 0.8135297298431396,
      "learning_rate": 0.00016283590130373882,
      "loss": 3.2192,
      "step": 67750
    },
    {
      "epoch": 7.29738456570875,
      "grad_norm": 0.8687593340873718,
      "learning_rate": 0.00016251266027367739,
      "loss": 3.2169,
      "step": 67800
    },
    {
      "epoch": 7.302766117748359,
      "grad_norm": 0.8100031018257141,
      "learning_rate": 0.00016218941924361595,
      "loss": 3.2052,
      "step": 67850
    },
    {
      "epoch": 7.308147669787967,
      "grad_norm": 0.7895699739456177,
      "learning_rate": 0.00016186617821355458,
      "loss": 3.2334,
      "step": 67900
    },
    {
      "epoch": 7.313529221827575,
      "grad_norm": 0.7895455956459045,
      "learning_rate": 0.00016154293718349314,
      "loss": 3.2231,
      "step": 67950
    },
    {
      "epoch": 7.3189107738671835,
      "grad_norm": 0.8295309543609619,
      "learning_rate": 0.00016121969615343174,
      "loss": 3.2006,
      "step": 68000
    },
    {
      "epoch": 7.3189107738671835,
      "eval_accuracy": 0.3882477342341672,
      "eval_loss": 3.351390838623047,
      "eval_runtime": 184.5883,
      "eval_samples_per_second": 97.574,
      "eval_steps_per_second": 6.1,
      "step": 68000
    },
    {
      "epoch": 7.324292325906791,
      "grad_norm": 0.8855544328689575,
      "learning_rate": 0.0001608964551233703,
      "loss": 3.214,
      "step": 68050
    },
    {
      "epoch": 7.3296738779464,
      "grad_norm": 0.8429924249649048,
      "learning_rate": 0.0001605732140933089,
      "loss": 3.2306,
      "step": 68100
    },
    {
      "epoch": 7.335055429986008,
      "grad_norm": 0.8152256011962891,
      "learning_rate": 0.0001602499730632475,
      "loss": 3.2301,
      "step": 68150
    },
    {
      "epoch": 7.340436982025616,
      "grad_norm": 0.8925390243530273,
      "learning_rate": 0.00015992673203318606,
      "loss": 3.2047,
      "step": 68200
    },
    {
      "epoch": 7.3458185340652244,
      "grad_norm": 0.8430379033088684,
      "learning_rate": 0.00015960349100312463,
      "loss": 3.2294,
      "step": 68250
    },
    {
      "epoch": 7.351200086104833,
      "grad_norm": 0.8633614778518677,
      "learning_rate": 0.00015928024997306325,
      "loss": 3.2202,
      "step": 68300
    },
    {
      "epoch": 7.356581638144441,
      "grad_norm": 0.8025208711624146,
      "learning_rate": 0.00015895700894300182,
      "loss": 3.2208,
      "step": 68350
    },
    {
      "epoch": 7.361963190184049,
      "grad_norm": 0.828989565372467,
      "learning_rate": 0.00015863376791294039,
      "loss": 3.2162,
      "step": 68400
    },
    {
      "epoch": 7.367344742223658,
      "grad_norm": 0.8749890923500061,
      "learning_rate": 0.000158310526882879,
      "loss": 3.2363,
      "step": 68450
    },
    {
      "epoch": 7.372726294263265,
      "grad_norm": 0.8234397768974304,
      "learning_rate": 0.00015798728585281758,
      "loss": 3.2358,
      "step": 68500
    },
    {
      "epoch": 7.378107846302874,
      "grad_norm": 0.8228161931037903,
      "learning_rate": 0.00015766404482275614,
      "loss": 3.2272,
      "step": 68550
    },
    {
      "epoch": 7.383489398342482,
      "grad_norm": 0.8303384780883789,
      "learning_rate": 0.00015734080379269474,
      "loss": 3.2314,
      "step": 68600
    },
    {
      "epoch": 7.38887095038209,
      "grad_norm": 0.8005957007408142,
      "learning_rate": 0.00015701756276263333,
      "loss": 3.2291,
      "step": 68650
    },
    {
      "epoch": 7.394252502421699,
      "grad_norm": 0.8635926842689514,
      "learning_rate": 0.0001566943217325719,
      "loss": 3.2199,
      "step": 68700
    },
    {
      "epoch": 7.399634054461306,
      "grad_norm": 0.7820467352867126,
      "learning_rate": 0.0001563710807025105,
      "loss": 3.2154,
      "step": 68750
    },
    {
      "epoch": 7.405015606500915,
      "grad_norm": 0.8372554779052734,
      "learning_rate": 0.00015604783967244906,
      "loss": 3.246,
      "step": 68800
    },
    {
      "epoch": 7.4103971585405235,
      "grad_norm": 0.8614478707313538,
      "learning_rate": 0.00015572459864238763,
      "loss": 3.2373,
      "step": 68850
    },
    {
      "epoch": 7.415778710580131,
      "grad_norm": 0.7786409854888916,
      "learning_rate": 0.00015540135761232625,
      "loss": 3.2196,
      "step": 68900
    },
    {
      "epoch": 7.42116026261974,
      "grad_norm": 0.80152827501297,
      "learning_rate": 0.00015507811658226482,
      "loss": 3.2348,
      "step": 68950
    },
    {
      "epoch": 7.426541814659347,
      "grad_norm": 0.8448081612586975,
      "learning_rate": 0.00015475487555220344,
      "loss": 3.2225,
      "step": 69000
    },
    {
      "epoch": 7.426541814659347,
      "eval_accuracy": 0.38895625971560543,
      "eval_loss": 3.3472721576690674,
      "eval_runtime": 184.8693,
      "eval_samples_per_second": 97.426,
      "eval_steps_per_second": 6.091,
      "step": 69000
    },
    {
      "epoch": 7.431923366698956,
      "grad_norm": 0.7952991127967834,
      "learning_rate": 0.000154431634522142,
      "loss": 3.2336,
      "step": 69050
    },
    {
      "epoch": 7.4373049187385645,
      "grad_norm": 0.8257370591163635,
      "learning_rate": 0.00015410839349208058,
      "loss": 3.2352,
      "step": 69100
    },
    {
      "epoch": 7.442686470778172,
      "grad_norm": 0.7818418741226196,
      "learning_rate": 0.00015378515246201917,
      "loss": 3.2127,
      "step": 69150
    },
    {
      "epoch": 7.448068022817781,
      "grad_norm": 0.7481926083564758,
      "learning_rate": 0.00015346191143195774,
      "loss": 3.2302,
      "step": 69200
    },
    {
      "epoch": 7.453449574857389,
      "grad_norm": 0.7871079444885254,
      "learning_rate": 0.00015313867040189633,
      "loss": 3.2131,
      "step": 69250
    },
    {
      "epoch": 7.458831126896997,
      "grad_norm": 0.8238528370857239,
      "learning_rate": 0.00015281542937183493,
      "loss": 3.2312,
      "step": 69300
    },
    {
      "epoch": 7.4642126789366054,
      "grad_norm": 0.8295047283172607,
      "learning_rate": 0.0001524921883417735,
      "loss": 3.2295,
      "step": 69350
    },
    {
      "epoch": 7.469594230976213,
      "grad_norm": 0.8163410425186157,
      "learning_rate": 0.00015216894731171206,
      "loss": 3.2072,
      "step": 69400
    },
    {
      "epoch": 7.474975783015822,
      "grad_norm": 0.8015903830528259,
      "learning_rate": 0.00015184570628165068,
      "loss": 3.2306,
      "step": 69450
    },
    {
      "epoch": 7.48035733505543,
      "grad_norm": 0.8221493363380432,
      "learning_rate": 0.00015152246525158925,
      "loss": 3.2241,
      "step": 69500
    },
    {
      "epoch": 7.485738887095038,
      "grad_norm": 0.8415564298629761,
      "learning_rate": 0.00015119922422152782,
      "loss": 3.2263,
      "step": 69550
    },
    {
      "epoch": 7.491120439134646,
      "grad_norm": 0.8357605338096619,
      "learning_rate": 0.00015087598319146644,
      "loss": 3.2443,
      "step": 69600
    },
    {
      "epoch": 7.496501991174255,
      "grad_norm": 0.8570162057876587,
      "learning_rate": 0.000150552742161405,
      "loss": 3.2303,
      "step": 69650
    },
    {
      "epoch": 7.501883543213863,
      "grad_norm": 0.8457155227661133,
      "learning_rate": 0.00015022950113134358,
      "loss": 3.2263,
      "step": 69700
    },
    {
      "epoch": 7.507265095253471,
      "grad_norm": 0.8195136785507202,
      "learning_rate": 0.00014990626010128217,
      "loss": 3.246,
      "step": 69750
    },
    {
      "epoch": 7.51264664729308,
      "grad_norm": 0.9330834150314331,
      "learning_rate": 0.00014958301907122077,
      "loss": 3.2381,
      "step": 69800
    },
    {
      "epoch": 7.518028199332687,
      "grad_norm": 0.836077094078064,
      "learning_rate": 0.00014925977804115933,
      "loss": 3.2522,
      "step": 69850
    },
    {
      "epoch": 7.523409751372296,
      "grad_norm": 0.8031020760536194,
      "learning_rate": 0.00014893653701109793,
      "loss": 3.237,
      "step": 69900
    },
    {
      "epoch": 7.528791303411904,
      "grad_norm": 0.8233877420425415,
      "learning_rate": 0.0001486132959810365,
      "loss": 3.229,
      "step": 69950
    },
    {
      "epoch": 7.534172855451512,
      "grad_norm": 0.8217124342918396,
      "learning_rate": 0.0001482900549509751,
      "loss": 3.2258,
      "step": 70000
    },
    {
      "epoch": 7.534172855451512,
      "eval_accuracy": 0.3891678068881802,
      "eval_loss": 3.345587968826294,
      "eval_runtime": 184.9082,
      "eval_samples_per_second": 97.405,
      "eval_steps_per_second": 6.09,
      "step": 70000
    },
    {
      "epoch": 7.539554407491121,
      "grad_norm": 0.8223305344581604,
      "learning_rate": 0.00014796681392091368,
      "loss": 3.2386,
      "step": 70050
    },
    {
      "epoch": 7.544935959530728,
      "grad_norm": 0.8422094583511353,
      "learning_rate": 0.00014764357289085228,
      "loss": 3.2385,
      "step": 70100
    },
    {
      "epoch": 7.550317511570337,
      "grad_norm": 0.8534138798713684,
      "learning_rate": 0.00014732033186079085,
      "loss": 3.2214,
      "step": 70150
    },
    {
      "epoch": 7.5556990636099455,
      "grad_norm": 0.8637493252754211,
      "learning_rate": 0.00014699709083072944,
      "loss": 3.2428,
      "step": 70200
    },
    {
      "epoch": 7.561080615649553,
      "grad_norm": 0.831741452217102,
      "learning_rate": 0.00014667384980066804,
      "loss": 3.2154,
      "step": 70250
    },
    {
      "epoch": 7.566462167689162,
      "grad_norm": 0.8678784370422363,
      "learning_rate": 0.00014635707359120784,
      "loss": 3.2371,
      "step": 70300
    },
    {
      "epoch": 7.57184371972877,
      "grad_norm": 0.8292410373687744,
      "learning_rate": 0.0001460338325611464,
      "loss": 3.2373,
      "step": 70350
    },
    {
      "epoch": 7.577225271768378,
      "grad_norm": 0.7928509712219238,
      "learning_rate": 0.00014571059153108498,
      "loss": 3.2306,
      "step": 70400
    },
    {
      "epoch": 7.5826068238079865,
      "grad_norm": 0.8370128870010376,
      "learning_rate": 0.00014538735050102357,
      "loss": 3.2222,
      "step": 70450
    },
    {
      "epoch": 7.587988375847594,
      "grad_norm": 0.801749587059021,
      "learning_rate": 0.00014506410947096217,
      "loss": 3.2312,
      "step": 70500
    },
    {
      "epoch": 7.593369927887203,
      "grad_norm": 0.8322213292121887,
      "learning_rate": 0.00014474086844090076,
      "loss": 3.2293,
      "step": 70550
    },
    {
      "epoch": 7.598751479926811,
      "grad_norm": 0.8021792769432068,
      "learning_rate": 0.00014441762741083933,
      "loss": 3.2364,
      "step": 70600
    },
    {
      "epoch": 7.604133031966419,
      "grad_norm": 0.8352303504943848,
      "learning_rate": 0.00014409438638077792,
      "loss": 3.2124,
      "step": 70650
    },
    {
      "epoch": 7.609514584006027,
      "grad_norm": 0.850394070148468,
      "learning_rate": 0.00014377114535071652,
      "loss": 3.2298,
      "step": 70700
    },
    {
      "epoch": 7.614896136045635,
      "grad_norm": 0.8542826771736145,
      "learning_rate": 0.0001434479043206551,
      "loss": 3.2264,
      "step": 70750
    },
    {
      "epoch": 7.620277688085244,
      "grad_norm": 0.8750932216644287,
      "learning_rate": 0.00014312466329059368,
      "loss": 3.2408,
      "step": 70800
    },
    {
      "epoch": 7.625659240124852,
      "grad_norm": 0.8510986566543579,
      "learning_rate": 0.00014280142226053225,
      "loss": 3.21,
      "step": 70850
    },
    {
      "epoch": 7.63104079216446,
      "grad_norm": 0.8305497765541077,
      "learning_rate": 0.00014247818123047084,
      "loss": 3.2402,
      "step": 70900
    },
    {
      "epoch": 7.636422344204068,
      "grad_norm": 0.8636519908905029,
      "learning_rate": 0.0001421549402004094,
      "loss": 3.2141,
      "step": 70950
    },
    {
      "epoch": 7.641803896243677,
      "grad_norm": 0.871688723564148,
      "learning_rate": 0.000141831699170348,
      "loss": 3.2407,
      "step": 71000
    },
    {
      "epoch": 7.641803896243677,
      "eval_accuracy": 0.3893689233835248,
      "eval_loss": 3.3400113582611084,
      "eval_runtime": 185.1188,
      "eval_samples_per_second": 97.294,
      "eval_steps_per_second": 6.083,
      "step": 71000
    },
    {
      "epoch": 7.647185448283285,
      "grad_norm": 0.8559573888778687,
      "learning_rate": 0.0001415084581402866,
      "loss": 3.2194,
      "step": 71050
    },
    {
      "epoch": 7.652567000322893,
      "grad_norm": 0.8053033947944641,
      "learning_rate": 0.00014118521711022517,
      "loss": 3.2275,
      "step": 71100
    },
    {
      "epoch": 7.657948552362502,
      "grad_norm": 0.872286856174469,
      "learning_rate": 0.00014086197608016376,
      "loss": 3.2224,
      "step": 71150
    },
    {
      "epoch": 7.663330104402109,
      "grad_norm": 0.7740647792816162,
      "learning_rate": 0.00014053873505010236,
      "loss": 3.2056,
      "step": 71200
    },
    {
      "epoch": 7.668711656441718,
      "grad_norm": 0.858696699142456,
      "learning_rate": 0.00014021549402004092,
      "loss": 3.2429,
      "step": 71250
    },
    {
      "epoch": 7.674093208481326,
      "grad_norm": 0.825717031955719,
      "learning_rate": 0.00013989225298997952,
      "loss": 3.24,
      "step": 71300
    },
    {
      "epoch": 7.679474760520934,
      "grad_norm": 0.8377315402030945,
      "learning_rate": 0.00013956901195991811,
      "loss": 3.2464,
      "step": 71350
    },
    {
      "epoch": 7.684856312560543,
      "grad_norm": 0.8150701522827148,
      "learning_rate": 0.00013924577092985668,
      "loss": 3.2356,
      "step": 71400
    },
    {
      "epoch": 7.69023786460015,
      "grad_norm": 0.8369122743606567,
      "learning_rate": 0.00013892252989979528,
      "loss": 3.2431,
      "step": 71450
    },
    {
      "epoch": 7.695619416639759,
      "grad_norm": 0.8841993808746338,
      "learning_rate": 0.00013859928886973384,
      "loss": 3.2415,
      "step": 71500
    },
    {
      "epoch": 7.7010009686793675,
      "grad_norm": 0.8426228165626526,
      "learning_rate": 0.00013827604783967244,
      "loss": 3.2186,
      "step": 71550
    },
    {
      "epoch": 7.706382520718975,
      "grad_norm": 0.819513738155365,
      "learning_rate": 0.000137952806809611,
      "loss": 3.2303,
      "step": 71600
    },
    {
      "epoch": 7.711764072758584,
      "grad_norm": 0.8620902895927429,
      "learning_rate": 0.0001376295657795496,
      "loss": 3.247,
      "step": 71650
    },
    {
      "epoch": 7.717145624798192,
      "grad_norm": 0.8760289549827576,
      "learning_rate": 0.0001373063247494882,
      "loss": 3.2234,
      "step": 71700
    },
    {
      "epoch": 7.7225271768378,
      "grad_norm": 0.8032657504081726,
      "learning_rate": 0.00013698308371942676,
      "loss": 3.2231,
      "step": 71750
    },
    {
      "epoch": 7.727908728877408,
      "grad_norm": 0.8196533918380737,
      "learning_rate": 0.00013665984268936536,
      "loss": 3.2375,
      "step": 71800
    },
    {
      "epoch": 7.733290280917016,
      "grad_norm": 0.795331597328186,
      "learning_rate": 0.00013633660165930395,
      "loss": 3.2346,
      "step": 71850
    },
    {
      "epoch": 7.738671832956625,
      "grad_norm": 0.8460344076156616,
      "learning_rate": 0.00013601336062924255,
      "loss": 3.2446,
      "step": 71900
    },
    {
      "epoch": 7.744053384996233,
      "grad_norm": 0.862134575843811,
      "learning_rate": 0.00013569011959918111,
      "loss": 3.2438,
      "step": 71950
    },
    {
      "epoch": 7.749434937035841,
      "grad_norm": 0.8252710700035095,
      "learning_rate": 0.0001353668785691197,
      "loss": 3.228,
      "step": 72000
    },
    {
      "epoch": 7.749434937035841,
      "eval_accuracy": 0.3897550757468175,
      "eval_loss": 3.3368613719940186,
      "eval_runtime": 184.6906,
      "eval_samples_per_second": 97.52,
      "eval_steps_per_second": 6.097,
      "step": 72000
    },
    {
      "epoch": 7.754816489075449,
      "grad_norm": 0.8688374161720276,
      "learning_rate": 0.00013504363753905828,
      "loss": 3.232,
      "step": 72050
    },
    {
      "epoch": 7.760198041115058,
      "grad_norm": 0.8163436055183411,
      "learning_rate": 0.00013472039650899687,
      "loss": 3.2337,
      "step": 72100
    },
    {
      "epoch": 7.765579593154666,
      "grad_norm": 0.7999975085258484,
      "learning_rate": 0.00013439715547893544,
      "loss": 3.2324,
      "step": 72150
    },
    {
      "epoch": 7.770961145194274,
      "grad_norm": 0.8432734608650208,
      "learning_rate": 0.00013407391444887403,
      "loss": 3.2204,
      "step": 72200
    },
    {
      "epoch": 7.776342697233883,
      "grad_norm": 0.8608858585357666,
      "learning_rate": 0.0001337506734188126,
      "loss": 3.2419,
      "step": 72250
    },
    {
      "epoch": 7.78172424927349,
      "grad_norm": 0.8750514388084412,
      "learning_rate": 0.0001334274323887512,
      "loss": 3.2275,
      "step": 72300
    },
    {
      "epoch": 7.787105801313099,
      "grad_norm": 0.8133453726768494,
      "learning_rate": 0.0001331041913586898,
      "loss": 3.2265,
      "step": 72350
    },
    {
      "epoch": 7.792487353352707,
      "grad_norm": 0.8550586700439453,
      "learning_rate": 0.00013278095032862838,
      "loss": 3.2262,
      "step": 72400
    },
    {
      "epoch": 7.797868905392315,
      "grad_norm": 0.8691797256469727,
      "learning_rate": 0.00013245770929856695,
      "loss": 3.243,
      "step": 72450
    },
    {
      "epoch": 7.803250457431924,
      "grad_norm": 0.883783757686615,
      "learning_rate": 0.00013213446826850555,
      "loss": 3.2444,
      "step": 72500
    },
    {
      "epoch": 7.808632009471531,
      "grad_norm": 0.8034313917160034,
      "learning_rate": 0.00013181122723844411,
      "loss": 3.2254,
      "step": 72550
    },
    {
      "epoch": 7.81401356151114,
      "grad_norm": 0.830434262752533,
      "learning_rate": 0.0001314879862083827,
      "loss": 3.235,
      "step": 72600
    },
    {
      "epoch": 7.819395113550748,
      "grad_norm": 0.8352016806602478,
      "learning_rate": 0.00013116474517832128,
      "loss": 3.2293,
      "step": 72650
    },
    {
      "epoch": 7.824776665590356,
      "grad_norm": 0.8240856528282166,
      "learning_rate": 0.00013084150414825987,
      "loss": 3.2342,
      "step": 72700
    },
    {
      "epoch": 7.830158217629965,
      "grad_norm": 0.7969643473625183,
      "learning_rate": 0.00013051826311819844,
      "loss": 3.2313,
      "step": 72750
    },
    {
      "epoch": 7.835539769669572,
      "grad_norm": 0.8742566108703613,
      "learning_rate": 0.00013019502208813703,
      "loss": 3.2344,
      "step": 72800
    },
    {
      "epoch": 7.840921321709181,
      "grad_norm": 0.8692363500595093,
      "learning_rate": 0.00012987178105807563,
      "loss": 3.2332,
      "step": 72850
    },
    {
      "epoch": 7.846302873748789,
      "grad_norm": 0.8365722298622131,
      "learning_rate": 0.00012954854002801422,
      "loss": 3.2264,
      "step": 72900
    },
    {
      "epoch": 7.851684425788397,
      "grad_norm": 0.828915536403656,
      "learning_rate": 0.0001292252989979528,
      "loss": 3.2234,
      "step": 72950
    },
    {
      "epoch": 7.857065977828006,
      "grad_norm": 0.8495892882347107,
      "learning_rate": 0.00012890205796789139,
      "loss": 3.2466,
      "step": 73000
    },
    {
      "epoch": 7.857065977828006,
      "eval_accuracy": 0.3902502069565881,
      "eval_loss": 3.332028388977051,
      "eval_runtime": 184.8344,
      "eval_samples_per_second": 97.444,
      "eval_steps_per_second": 6.092,
      "step": 73000
    },
    {
      "epoch": 7.862447529867614,
      "grad_norm": 0.8604885339736938,
      "learning_rate": 0.00012857881693782998,
      "loss": 3.228,
      "step": 73050
    },
    {
      "epoch": 7.867829081907222,
      "grad_norm": 0.8913674354553223,
      "learning_rate": 0.00012825557590776855,
      "loss": 3.2537,
      "step": 73100
    },
    {
      "epoch": 7.87321063394683,
      "grad_norm": 0.8491577506065369,
      "learning_rate": 0.00012793233487770714,
      "loss": 3.222,
      "step": 73150
    },
    {
      "epoch": 7.878592185986438,
      "grad_norm": 0.9810826182365417,
      "learning_rate": 0.0001276090938476457,
      "loss": 3.2323,
      "step": 73200
    },
    {
      "epoch": 7.883973738026047,
      "grad_norm": 0.8327536582946777,
      "learning_rate": 0.0001272858528175843,
      "loss": 3.235,
      "step": 73250
    },
    {
      "epoch": 7.889355290065655,
      "grad_norm": 0.8036714196205139,
      "learning_rate": 0.00012696261178752287,
      "loss": 3.2236,
      "step": 73300
    },
    {
      "epoch": 7.894736842105263,
      "grad_norm": 0.8222293853759766,
      "learning_rate": 0.00012663937075746147,
      "loss": 3.2249,
      "step": 73350
    },
    {
      "epoch": 7.900118394144871,
      "grad_norm": 0.8565043210983276,
      "learning_rate": 0.00012631612972740006,
      "loss": 3.2353,
      "step": 73400
    },
    {
      "epoch": 7.90549994618448,
      "grad_norm": 0.8443336486816406,
      "learning_rate": 0.00012599288869733863,
      "loss": 3.2214,
      "step": 73450
    },
    {
      "epoch": 7.910881498224088,
      "grad_norm": 0.9020589590072632,
      "learning_rate": 0.00012566964766727722,
      "loss": 3.2482,
      "step": 73500
    },
    {
      "epoch": 7.916263050263696,
      "grad_norm": 0.8561992049217224,
      "learning_rate": 0.00012534640663721582,
      "loss": 3.2166,
      "step": 73550
    },
    {
      "epoch": 7.921644602303305,
      "grad_norm": 0.8745976090431213,
      "learning_rate": 0.00012502316560715439,
      "loss": 3.2524,
      "step": 73600
    },
    {
      "epoch": 7.927026154342912,
      "grad_norm": 0.8272269368171692,
      "learning_rate": 0.00012469992457709298,
      "loss": 3.2354,
      "step": 73650
    },
    {
      "epoch": 7.932407706382521,
      "grad_norm": 0.9126244783401489,
      "learning_rate": 0.00012437668354703158,
      "loss": 3.2354,
      "step": 73700
    },
    {
      "epoch": 7.937789258422129,
      "grad_norm": 0.8359974026679993,
      "learning_rate": 0.00012405344251697014,
      "loss": 3.2412,
      "step": 73750
    },
    {
      "epoch": 7.943170810461737,
      "grad_norm": 0.8640374541282654,
      "learning_rate": 0.00012373020148690874,
      "loss": 3.2426,
      "step": 73800
    },
    {
      "epoch": 7.948552362501346,
      "grad_norm": 0.878259003162384,
      "learning_rate": 0.0001234069604568473,
      "loss": 3.223,
      "step": 73850
    },
    {
      "epoch": 7.953933914540953,
      "grad_norm": 0.9081090688705444,
      "learning_rate": 0.0001230837194267859,
      "loss": 3.2481,
      "step": 73900
    },
    {
      "epoch": 7.959315466580562,
      "grad_norm": 0.8142951130867004,
      "learning_rate": 0.00012276047839672447,
      "loss": 3.2287,
      "step": 73950
    },
    {
      "epoch": 7.96469701862017,
      "grad_norm": 0.8311217427253723,
      "learning_rate": 0.00012243723736666306,
      "loss": 3.2126,
      "step": 74000
    },
    {
      "epoch": 7.96469701862017,
      "eval_accuracy": 0.39052531606853436,
      "eval_loss": 3.3307554721832275,
      "eval_runtime": 184.6893,
      "eval_samples_per_second": 97.521,
      "eval_steps_per_second": 6.097,
      "step": 74000
    },
    {
      "epoch": 7.970078570659778,
      "grad_norm": 0.8098016381263733,
      "learning_rate": 0.00012211399633660166,
      "loss": 3.2278,
      "step": 74050
    },
    {
      "epoch": 7.975460122699387,
      "grad_norm": 0.8477680683135986,
      "learning_rate": 0.00012179075530654022,
      "loss": 3.2304,
      "step": 74100
    },
    {
      "epoch": 7.980841674738995,
      "grad_norm": 0.825594961643219,
      "learning_rate": 0.00012146751427647882,
      "loss": 3.2386,
      "step": 74150
    },
    {
      "epoch": 7.986223226778603,
      "grad_norm": 0.8377440571784973,
      "learning_rate": 0.0001211442732464174,
      "loss": 3.2335,
      "step": 74200
    },
    {
      "epoch": 7.991604778818211,
      "grad_norm": 0.8370752930641174,
      "learning_rate": 0.00012082103221635598,
      "loss": 3.2272,
      "step": 74250
    },
    {
      "epoch": 7.996986330857819,
      "grad_norm": 0.8993654251098633,
      "learning_rate": 0.00012049779118629456,
      "loss": 3.2465,
      "step": 74300
    },
    {
      "epoch": 8.002367882897428,
      "grad_norm": 0.8433480858802795,
      "learning_rate": 0.0001201810149768344,
      "loss": 3.1976,
      "step": 74350
    },
    {
      "epoch": 8.007749434937036,
      "grad_norm": 0.8284233808517456,
      "learning_rate": 0.00011985777394677296,
      "loss": 3.1514,
      "step": 74400
    },
    {
      "epoch": 8.013130986976645,
      "grad_norm": 0.8333739638328552,
      "learning_rate": 0.00011953453291671154,
      "loss": 3.1506,
      "step": 74450
    },
    {
      "epoch": 8.018512539016251,
      "grad_norm": 0.8498620986938477,
      "learning_rate": 0.00011921129188665014,
      "loss": 3.1347,
      "step": 74500
    },
    {
      "epoch": 8.02389409105586,
      "grad_norm": 0.8060201406478882,
      "learning_rate": 0.0001188880508565887,
      "loss": 3.1446,
      "step": 74550
    },
    {
      "epoch": 8.029275643095469,
      "grad_norm": 0.8369296789169312,
      "learning_rate": 0.0001185648098265273,
      "loss": 3.1747,
      "step": 74600
    },
    {
      "epoch": 8.034657195135077,
      "grad_norm": 0.8029821515083313,
      "learning_rate": 0.0001182415687964659,
      "loss": 3.1674,
      "step": 74650
    },
    {
      "epoch": 8.040038747174686,
      "grad_norm": 0.8536189198493958,
      "learning_rate": 0.00011791832776640448,
      "loss": 3.1492,
      "step": 74700
    },
    {
      "epoch": 8.045420299214294,
      "grad_norm": 0.8600338101387024,
      "learning_rate": 0.00011759508673634306,
      "loss": 3.1523,
      "step": 74750
    },
    {
      "epoch": 8.050801851253901,
      "grad_norm": 0.838543713092804,
      "learning_rate": 0.00011727184570628164,
      "loss": 3.1805,
      "step": 74800
    },
    {
      "epoch": 8.05618340329351,
      "grad_norm": 0.7800434827804565,
      "learning_rate": 0.00011694860467622023,
      "loss": 3.172,
      "step": 74850
    },
    {
      "epoch": 8.061564955333118,
      "grad_norm": 0.8202197551727295,
      "learning_rate": 0.0001166253636461588,
      "loss": 3.1607,
      "step": 74900
    },
    {
      "epoch": 8.066946507372727,
      "grad_norm": 0.8486565947532654,
      "learning_rate": 0.0001163021226160974,
      "loss": 3.1748,
      "step": 74950
    },
    {
      "epoch": 8.072328059412335,
      "grad_norm": 0.9214633107185364,
      "learning_rate": 0.00011597888158603598,
      "loss": 3.1488,
      "step": 75000
    },
    {
      "epoch": 8.072328059412335,
      "eval_accuracy": 0.3903842846201511,
      "eval_loss": 3.3340110778808594,
      "eval_runtime": 185.0586,
      "eval_samples_per_second": 97.326,
      "eval_steps_per_second": 6.085,
      "step": 75000
    },
    {
      "epoch": 8.077709611451942,
      "grad_norm": 0.868916392326355,
      "learning_rate": 0.00011565564055597456,
      "loss": 3.1697,
      "step": 75050
    },
    {
      "epoch": 8.08309116349155,
      "grad_norm": 0.8408175706863403,
      "learning_rate": 0.00011533239952591314,
      "loss": 3.1719,
      "step": 75100
    },
    {
      "epoch": 8.088472715531159,
      "grad_norm": 0.8719428777694702,
      "learning_rate": 0.00011500915849585173,
      "loss": 3.182,
      "step": 75150
    },
    {
      "epoch": 8.093854267570768,
      "grad_norm": 0.7724485397338867,
      "learning_rate": 0.00011468591746579033,
      "loss": 3.1634,
      "step": 75200
    },
    {
      "epoch": 8.099235819610376,
      "grad_norm": 0.8824598789215088,
      "learning_rate": 0.0001143626764357289,
      "loss": 3.1758,
      "step": 75250
    },
    {
      "epoch": 8.104617371649983,
      "grad_norm": 0.8987236022949219,
      "learning_rate": 0.00011403943540566749,
      "loss": 3.1593,
      "step": 75300
    },
    {
      "epoch": 8.109998923689592,
      "grad_norm": 0.8450270295143127,
      "learning_rate": 0.00011371619437560607,
      "loss": 3.164,
      "step": 75350
    },
    {
      "epoch": 8.1153804757292,
      "grad_norm": 0.8858851790428162,
      "learning_rate": 0.00011339295334554464,
      "loss": 3.1648,
      "step": 75400
    },
    {
      "epoch": 8.120762027768809,
      "grad_norm": 0.83014315366745,
      "learning_rate": 0.00011306971231548323,
      "loss": 3.1557,
      "step": 75450
    },
    {
      "epoch": 8.126143579808417,
      "grad_norm": 0.8151319622993469,
      "learning_rate": 0.00011274647128542183,
      "loss": 3.1569,
      "step": 75500
    },
    {
      "epoch": 8.131525131848026,
      "grad_norm": 0.8844127655029297,
      "learning_rate": 0.0001124232302553604,
      "loss": 3.1876,
      "step": 75550
    },
    {
      "epoch": 8.136906683887632,
      "grad_norm": 0.8928391933441162,
      "learning_rate": 0.00011209998922529899,
      "loss": 3.1632,
      "step": 75600
    },
    {
      "epoch": 8.142288235927241,
      "grad_norm": 0.844195544719696,
      "learning_rate": 0.00011177674819523757,
      "loss": 3.1593,
      "step": 75650
    },
    {
      "epoch": 8.14766978796685,
      "grad_norm": 0.8816542625427246,
      "learning_rate": 0.00011145350716517617,
      "loss": 3.1543,
      "step": 75700
    },
    {
      "epoch": 8.153051340006458,
      "grad_norm": 0.917953372001648,
      "learning_rate": 0.00011113026613511473,
      "loss": 3.1802,
      "step": 75750
    },
    {
      "epoch": 8.158432892046067,
      "grad_norm": 0.8676404356956482,
      "learning_rate": 0.00011081348992565455,
      "loss": 3.1598,
      "step": 75800
    },
    {
      "epoch": 8.163814444085673,
      "grad_norm": 0.8764878511428833,
      "learning_rate": 0.00011049024889559314,
      "loss": 3.1617,
      "step": 75850
    },
    {
      "epoch": 8.169195996125282,
      "grad_norm": 0.8242963552474976,
      "learning_rate": 0.00011016700786553172,
      "loss": 3.1673,
      "step": 75900
    },
    {
      "epoch": 8.17457754816489,
      "grad_norm": 0.8988597989082336,
      "learning_rate": 0.00010984376683547031,
      "loss": 3.181,
      "step": 75950
    },
    {
      "epoch": 8.1799591002045,
      "grad_norm": 0.8427569270133972,
      "learning_rate": 0.00010952052580540889,
      "loss": 3.1626,
      "step": 76000
    },
    {
      "epoch": 8.1799591002045,
      "eval_accuracy": 0.3904491503941764,
      "eval_loss": 3.3352701663970947,
      "eval_runtime": 184.661,
      "eval_samples_per_second": 97.535,
      "eval_steps_per_second": 6.098,
      "step": 76000
    },
    {
      "epoch": 8.185340652244108,
      "grad_norm": 0.8113105297088623,
      "learning_rate": 0.00010919728477534747,
      "loss": 3.1803,
      "step": 76050
    },
    {
      "epoch": 8.190722204283716,
      "grad_norm": 0.8857055902481079,
      "learning_rate": 0.00010887404374528605,
      "loss": 3.1951,
      "step": 76100
    },
    {
      "epoch": 8.196103756323323,
      "grad_norm": 0.8819203972816467,
      "learning_rate": 0.00010855080271522465,
      "loss": 3.157,
      "step": 76150
    },
    {
      "epoch": 8.201485308362932,
      "grad_norm": 0.920045793056488,
      "learning_rate": 0.00010822756168516322,
      "loss": 3.1773,
      "step": 76200
    },
    {
      "epoch": 8.20686686040254,
      "grad_norm": 0.9266934990882874,
      "learning_rate": 0.00010790432065510181,
      "loss": 3.18,
      "step": 76250
    },
    {
      "epoch": 8.212248412442149,
      "grad_norm": 0.8361415266990662,
      "learning_rate": 0.0001075810796250404,
      "loss": 3.1652,
      "step": 76300
    },
    {
      "epoch": 8.217629964481757,
      "grad_norm": 0.9125852584838867,
      "learning_rate": 0.00010725783859497897,
      "loss": 3.1847,
      "step": 76350
    },
    {
      "epoch": 8.223011516521364,
      "grad_norm": 0.8211091160774231,
      "learning_rate": 0.00010693459756491757,
      "loss": 3.1481,
      "step": 76400
    },
    {
      "epoch": 8.228393068560973,
      "grad_norm": 0.8206071257591248,
      "learning_rate": 0.00010661135653485615,
      "loss": 3.1628,
      "step": 76450
    },
    {
      "epoch": 8.233774620600581,
      "grad_norm": 0.8912805318832397,
      "learning_rate": 0.00010628811550479474,
      "loss": 3.1706,
      "step": 76500
    },
    {
      "epoch": 8.23915617264019,
      "grad_norm": 0.8347760438919067,
      "learning_rate": 0.00010596487447473331,
      "loss": 3.1822,
      "step": 76550
    },
    {
      "epoch": 8.244537724679798,
      "grad_norm": 0.8456274271011353,
      "learning_rate": 0.0001056416334446719,
      "loss": 3.1729,
      "step": 76600
    },
    {
      "epoch": 8.249919276719407,
      "grad_norm": 0.8533112406730652,
      "learning_rate": 0.00010531839241461049,
      "loss": 3.1695,
      "step": 76650
    },
    {
      "epoch": 8.255300828759013,
      "grad_norm": 0.8624753952026367,
      "learning_rate": 0.00010499515138454907,
      "loss": 3.1648,
      "step": 76700
    },
    {
      "epoch": 8.260682380798622,
      "grad_norm": 0.8519622683525085,
      "learning_rate": 0.00010467191035448765,
      "loss": 3.1798,
      "step": 76750
    },
    {
      "epoch": 8.26606393283823,
      "grad_norm": 0.838147759437561,
      "learning_rate": 0.00010434866932442624,
      "loss": 3.1683,
      "step": 76800
    },
    {
      "epoch": 8.27144548487784,
      "grad_norm": 0.903648853302002,
      "learning_rate": 0.00010402542829436481,
      "loss": 3.1834,
      "step": 76850
    },
    {
      "epoch": 8.276827036917448,
      "grad_norm": 0.8623703718185425,
      "learning_rate": 0.0001037021872643034,
      "loss": 3.1707,
      "step": 76900
    },
    {
      "epoch": 8.282208588957054,
      "grad_norm": 0.8603115081787109,
      "learning_rate": 0.00010337894623424199,
      "loss": 3.1824,
      "step": 76950
    },
    {
      "epoch": 8.287590140996663,
      "grad_norm": 0.8683320879936218,
      "learning_rate": 0.00010305570520418058,
      "loss": 3.1706,
      "step": 77000
    },
    {
      "epoch": 8.287590140996663,
      "eval_accuracy": 0.39092265967927187,
      "eval_loss": 3.3312289714813232,
      "eval_runtime": 184.7891,
      "eval_samples_per_second": 97.468,
      "eval_steps_per_second": 6.093,
      "step": 77000
    },
    {
      "epoch": 8.292971693036272,
      "grad_norm": 0.8868846893310547,
      "learning_rate": 0.00010273246417411915,
      "loss": 3.1657,
      "step": 77050
    },
    {
      "epoch": 8.29835324507588,
      "grad_norm": 0.835442304611206,
      "learning_rate": 0.00010240922314405774,
      "loss": 3.1762,
      "step": 77100
    },
    {
      "epoch": 8.303734797115489,
      "grad_norm": 0.8506765961647034,
      "learning_rate": 0.00010208598211399634,
      "loss": 3.1836,
      "step": 77150
    },
    {
      "epoch": 8.309116349155097,
      "grad_norm": 0.9037638306617737,
      "learning_rate": 0.0001017627410839349,
      "loss": 3.1595,
      "step": 77200
    },
    {
      "epoch": 8.314497901194704,
      "grad_norm": 0.8438506722450256,
      "learning_rate": 0.0001014395000538735,
      "loss": 3.1774,
      "step": 77250
    },
    {
      "epoch": 8.319879453234313,
      "grad_norm": 0.8813534379005432,
      "learning_rate": 0.00010111625902381208,
      "loss": 3.1947,
      "step": 77300
    },
    {
      "epoch": 8.325261005273921,
      "grad_norm": 0.8885173797607422,
      "learning_rate": 0.00010079301799375066,
      "loss": 3.1824,
      "step": 77350
    },
    {
      "epoch": 8.33064255731353,
      "grad_norm": 0.8622186183929443,
      "learning_rate": 0.00010046977696368924,
      "loss": 3.1645,
      "step": 77400
    },
    {
      "epoch": 8.336024109353138,
      "grad_norm": 0.891852080821991,
      "learning_rate": 0.00010014653593362784,
      "loss": 3.1819,
      "step": 77450
    },
    {
      "epoch": 8.341405661392745,
      "grad_norm": 0.8418218493461609,
      "learning_rate": 9.982329490356642e-05,
      "loss": 3.1684,
      "step": 77500
    },
    {
      "epoch": 8.346787213432354,
      "grad_norm": 0.8554936051368713,
      "learning_rate": 9.9500053873505e-05,
      "loss": 3.1878,
      "step": 77550
    },
    {
      "epoch": 8.352168765471962,
      "grad_norm": 0.854648232460022,
      "learning_rate": 9.917681284344358e-05,
      "loss": 3.1779,
      "step": 77600
    },
    {
      "epoch": 8.35755031751157,
      "grad_norm": 0.8960241079330444,
      "learning_rate": 9.885357181338218e-05,
      "loss": 3.1708,
      "step": 77650
    },
    {
      "epoch": 8.36293186955118,
      "grad_norm": 0.9255419969558716,
      "learning_rate": 9.853033078332074e-05,
      "loss": 3.1909,
      "step": 77700
    },
    {
      "epoch": 8.368313421590786,
      "grad_norm": 0.8990132808685303,
      "learning_rate": 9.820708975325934e-05,
      "loss": 3.1713,
      "step": 77750
    },
    {
      "epoch": 8.373694973630395,
      "grad_norm": 0.85789954662323,
      "learning_rate": 9.788384872319793e-05,
      "loss": 3.1605,
      "step": 77800
    },
    {
      "epoch": 8.379076525670003,
      "grad_norm": 0.858098030090332,
      "learning_rate": 9.75606076931365e-05,
      "loss": 3.1687,
      "step": 77850
    },
    {
      "epoch": 8.384458077709612,
      "grad_norm": 0.8912584781646729,
      "learning_rate": 9.723736666307508e-05,
      "loss": 3.1633,
      "step": 77900
    },
    {
      "epoch": 8.38983962974922,
      "grad_norm": 0.8760279417037964,
      "learning_rate": 9.691412563301368e-05,
      "loss": 3.163,
      "step": 77950
    },
    {
      "epoch": 8.395221181788829,
      "grad_norm": 0.8484280705451965,
      "learning_rate": 9.659088460295227e-05,
      "loss": 3.1875,
      "step": 78000
    },
    {
      "epoch": 8.395221181788829,
      "eval_accuracy": 0.39122645315360116,
      "eval_loss": 3.326995372772217,
      "eval_runtime": 184.8092,
      "eval_samples_per_second": 97.457,
      "eval_steps_per_second": 6.093,
      "step": 78000
    },
    {
      "epoch": 8.400602733828435,
      "grad_norm": 0.8364370465278625,
      "learning_rate": 9.626764357289084e-05,
      "loss": 3.1871,
      "step": 78050
    },
    {
      "epoch": 8.405984285868044,
      "grad_norm": 0.8873591423034668,
      "learning_rate": 9.594440254282943e-05,
      "loss": 3.1791,
      "step": 78100
    },
    {
      "epoch": 8.411365837907653,
      "grad_norm": 0.8371385335922241,
      "learning_rate": 9.562116151276802e-05,
      "loss": 3.1628,
      "step": 78150
    },
    {
      "epoch": 8.416747389947261,
      "grad_norm": 0.8212653994560242,
      "learning_rate": 9.52979204827066e-05,
      "loss": 3.1585,
      "step": 78200
    },
    {
      "epoch": 8.42212894198687,
      "grad_norm": 0.8428384065628052,
      "learning_rate": 9.497467945264518e-05,
      "loss": 3.2135,
      "step": 78250
    },
    {
      "epoch": 8.427510494026476,
      "grad_norm": 0.8416067361831665,
      "learning_rate": 9.465143842258377e-05,
      "loss": 3.1811,
      "step": 78300
    },
    {
      "epoch": 8.432892046066085,
      "grad_norm": 0.8841109871864319,
      "learning_rate": 9.432819739252234e-05,
      "loss": 3.1834,
      "step": 78350
    },
    {
      "epoch": 8.438273598105694,
      "grad_norm": 0.8793404698371887,
      "learning_rate": 9.400495636246093e-05,
      "loss": 3.1823,
      "step": 78400
    },
    {
      "epoch": 8.443655150145302,
      "grad_norm": 0.8999808430671692,
      "learning_rate": 9.368171533239952e-05,
      "loss": 3.1813,
      "step": 78450
    },
    {
      "epoch": 8.44903670218491,
      "grad_norm": 0.832278847694397,
      "learning_rate": 9.335847430233811e-05,
      "loss": 3.1914,
      "step": 78500
    },
    {
      "epoch": 8.45441825422452,
      "grad_norm": 0.8846303820610046,
      "learning_rate": 9.303523327227668e-05,
      "loss": 3.1763,
      "step": 78550
    },
    {
      "epoch": 8.459799806264126,
      "grad_norm": 0.8395057320594788,
      "learning_rate": 9.271199224221527e-05,
      "loss": 3.1892,
      "step": 78600
    },
    {
      "epoch": 8.465181358303735,
      "grad_norm": 0.8631083369255066,
      "learning_rate": 9.238875121215387e-05,
      "loss": 3.1991,
      "step": 78650
    },
    {
      "epoch": 8.470562910343343,
      "grad_norm": 0.8669479489326477,
      "learning_rate": 9.206551018209243e-05,
      "loss": 3.1691,
      "step": 78700
    },
    {
      "epoch": 8.475944462382952,
      "grad_norm": 0.8604099154472351,
      "learning_rate": 9.174226915203103e-05,
      "loss": 3.1691,
      "step": 78750
    },
    {
      "epoch": 8.48132601442256,
      "grad_norm": 0.8545402884483337,
      "learning_rate": 9.141902812196961e-05,
      "loss": 3.1751,
      "step": 78800
    },
    {
      "epoch": 8.486707566462167,
      "grad_norm": 0.8492422699928284,
      "learning_rate": 9.109578709190818e-05,
      "loss": 3.1837,
      "step": 78850
    },
    {
      "epoch": 8.492089118501776,
      "grad_norm": 0.8209715485572815,
      "learning_rate": 9.077254606184677e-05,
      "loss": 3.1882,
      "step": 78900
    },
    {
      "epoch": 8.497470670541384,
      "grad_norm": 0.876168429851532,
      "learning_rate": 9.044930503178537e-05,
      "loss": 3.1854,
      "step": 78950
    },
    {
      "epoch": 8.502852222580993,
      "grad_norm": 0.9051874876022339,
      "learning_rate": 9.012606400172395e-05,
      "loss": 3.1783,
      "step": 79000
    },
    {
      "epoch": 8.502852222580993,
      "eval_accuracy": 0.3914579924575338,
      "eval_loss": 3.3256032466888428,
      "eval_runtime": 185.3855,
      "eval_samples_per_second": 97.154,
      "eval_steps_per_second": 6.074,
      "step": 79000
    },
    {
      "epoch": 8.508233774620601,
      "grad_norm": 0.9354413747787476,
      "learning_rate": 8.980282297166253e-05,
      "loss": 3.1785,
      "step": 79050
    },
    {
      "epoch": 8.513615326660208,
      "grad_norm": 0.8846203684806824,
      "learning_rate": 8.947958194160111e-05,
      "loss": 3.2052,
      "step": 79100
    },
    {
      "epoch": 8.518996878699816,
      "grad_norm": 0.8579129576683044,
      "learning_rate": 8.91563409115397e-05,
      "loss": 3.1818,
      "step": 79150
    },
    {
      "epoch": 8.524378430739425,
      "grad_norm": 0.8579406142234802,
      "learning_rate": 8.883309988147827e-05,
      "loss": 3.1827,
      "step": 79200
    },
    {
      "epoch": 8.529759982779034,
      "grad_norm": 0.8814583420753479,
      "learning_rate": 8.850985885141687e-05,
      "loss": 3.1749,
      "step": 79250
    },
    {
      "epoch": 8.535141534818642,
      "grad_norm": 0.8940314054489136,
      "learning_rate": 8.818661782135545e-05,
      "loss": 3.1611,
      "step": 79300
    },
    {
      "epoch": 8.54052308685825,
      "grad_norm": 0.9121882319450378,
      "learning_rate": 8.786337679129403e-05,
      "loss": 3.1679,
      "step": 79350
    },
    {
      "epoch": 8.545904638897857,
      "grad_norm": 0.9019620418548584,
      "learning_rate": 8.754013576123261e-05,
      "loss": 3.1741,
      "step": 79400
    },
    {
      "epoch": 8.551286190937466,
      "grad_norm": 0.9139887690544128,
      "learning_rate": 8.72168947311712e-05,
      "loss": 3.1732,
      "step": 79450
    },
    {
      "epoch": 8.556667742977075,
      "grad_norm": 0.8666207790374756,
      "learning_rate": 8.68936537011098e-05,
      "loss": 3.1821,
      "step": 79500
    },
    {
      "epoch": 8.562049295016683,
      "grad_norm": 0.8577825427055359,
      "learning_rate": 8.657041267104837e-05,
      "loss": 3.1826,
      "step": 79550
    },
    {
      "epoch": 8.567430847056292,
      "grad_norm": 0.8668467402458191,
      "learning_rate": 8.624717164098696e-05,
      "loss": 3.1725,
      "step": 79600
    },
    {
      "epoch": 8.572812399095898,
      "grad_norm": 0.8373755216598511,
      "learning_rate": 8.592393061092554e-05,
      "loss": 3.1797,
      "step": 79650
    },
    {
      "epoch": 8.578193951135507,
      "grad_norm": 0.8314027786254883,
      "learning_rate": 8.560068958086412e-05,
      "loss": 3.1849,
      "step": 79700
    },
    {
      "epoch": 8.583575503175116,
      "grad_norm": 0.9058263897895813,
      "learning_rate": 8.52774485508027e-05,
      "loss": 3.176,
      "step": 79750
    },
    {
      "epoch": 8.588957055214724,
      "grad_norm": 0.9133097529411316,
      "learning_rate": 8.49542075207413e-05,
      "loss": 3.1968,
      "step": 79800
    },
    {
      "epoch": 8.594338607254333,
      "grad_norm": 0.8769819140434265,
      "learning_rate": 8.46374313112811e-05,
      "loss": 3.1829,
      "step": 79850
    },
    {
      "epoch": 8.599720159293941,
      "grad_norm": 1.0664840936660767,
      "learning_rate": 8.431419028121969e-05,
      "loss": 3.1828,
      "step": 79900
    },
    {
      "epoch": 8.605101711333548,
      "grad_norm": 0.8446348905563354,
      "learning_rate": 8.399094925115828e-05,
      "loss": 3.1563,
      "step": 79950
    },
    {
      "epoch": 8.610483263373157,
      "grad_norm": 0.8764998912811279,
      "learning_rate": 8.366770822109685e-05,
      "loss": 3.1571,
      "step": 80000
    },
    {
      "epoch": 8.610483263373157,
      "eval_accuracy": 0.39208394174423306,
      "eval_loss": 3.321202039718628,
      "eval_runtime": 184.6735,
      "eval_samples_per_second": 97.529,
      "eval_steps_per_second": 6.097,
      "step": 80000
    },
    {
      "epoch": 8.615864815412765,
      "grad_norm": 0.8480128049850464,
      "learning_rate": 8.334446719103544e-05,
      "loss": 3.189,
      "step": 80050
    },
    {
      "epoch": 8.621246367452374,
      "grad_norm": 0.8969175219535828,
      "learning_rate": 8.302122616097403e-05,
      "loss": 3.1913,
      "step": 80100
    },
    {
      "epoch": 8.626627919491982,
      "grad_norm": 0.8639907240867615,
      "learning_rate": 8.269798513091261e-05,
      "loss": 3.1644,
      "step": 80150
    },
    {
      "epoch": 8.632009471531589,
      "grad_norm": 0.8794361352920532,
      "learning_rate": 8.237474410085119e-05,
      "loss": 3.1668,
      "step": 80200
    },
    {
      "epoch": 8.637391023571197,
      "grad_norm": 0.885933518409729,
      "learning_rate": 8.205150307078978e-05,
      "loss": 3.1819,
      "step": 80250
    },
    {
      "epoch": 8.642772575610806,
      "grad_norm": 0.8281596302986145,
      "learning_rate": 8.172826204072838e-05,
      "loss": 3.1691,
      "step": 80300
    },
    {
      "epoch": 8.648154127650415,
      "grad_norm": 0.9188771843910217,
      "learning_rate": 8.140502101066694e-05,
      "loss": 3.1883,
      "step": 80350
    },
    {
      "epoch": 8.653535679690023,
      "grad_norm": 0.9139700531959534,
      "learning_rate": 8.108177998060553e-05,
      "loss": 3.1712,
      "step": 80400
    },
    {
      "epoch": 8.658917231729632,
      "grad_norm": 0.8853664398193359,
      "learning_rate": 8.075853895054412e-05,
      "loss": 3.1819,
      "step": 80450
    },
    {
      "epoch": 8.664298783769238,
      "grad_norm": 0.8718395233154297,
      "learning_rate": 8.043529792048269e-05,
      "loss": 3.1864,
      "step": 80500
    },
    {
      "epoch": 8.669680335808847,
      "grad_norm": 0.8565669655799866,
      "learning_rate": 8.011205689042128e-05,
      "loss": 3.1692,
      "step": 80550
    },
    {
      "epoch": 8.675061887848456,
      "grad_norm": 0.8765687942504883,
      "learning_rate": 7.978881586035988e-05,
      "loss": 3.2026,
      "step": 80600
    },
    {
      "epoch": 8.680443439888064,
      "grad_norm": 0.8425050973892212,
      "learning_rate": 7.946557483029845e-05,
      "loss": 3.1826,
      "step": 80650
    },
    {
      "epoch": 8.685824991927673,
      "grad_norm": 0.8891403079032898,
      "learning_rate": 7.914233380023704e-05,
      "loss": 3.1744,
      "step": 80700
    },
    {
      "epoch": 8.69120654396728,
      "grad_norm": 0.8497437834739685,
      "learning_rate": 7.881909277017562e-05,
      "loss": 3.1793,
      "step": 80750
    },
    {
      "epoch": 8.696588096006888,
      "grad_norm": 0.8760660886764526,
      "learning_rate": 7.849585174011422e-05,
      "loss": 3.1883,
      "step": 80800
    },
    {
      "epoch": 8.701969648046497,
      "grad_norm": 0.9156056642532349,
      "learning_rate": 7.817261071005278e-05,
      "loss": 3.1886,
      "step": 80850
    },
    {
      "epoch": 8.707351200086105,
      "grad_norm": 0.9590579867362976,
      "learning_rate": 7.784936967999138e-05,
      "loss": 3.1879,
      "step": 80900
    },
    {
      "epoch": 8.712732752125714,
      "grad_norm": 0.8654018044471741,
      "learning_rate": 7.752612864992996e-05,
      "loss": 3.1743,
      "step": 80950
    },
    {
      "epoch": 8.718114304165322,
      "grad_norm": 1.0002062320709229,
      "learning_rate": 7.720288761986854e-05,
      "loss": 3.2066,
      "step": 81000
    },
    {
      "epoch": 8.718114304165322,
      "eval_accuracy": 0.39225344024922365,
      "eval_loss": 3.317610025405884,
      "eval_runtime": 184.756,
      "eval_samples_per_second": 97.485,
      "eval_steps_per_second": 6.095,
      "step": 81000
    },
    {
      "epoch": 8.723495856204929,
      "grad_norm": 0.8658928871154785,
      "learning_rate": 7.687964658980712e-05,
      "loss": 3.1837,
      "step": 81050
    },
    {
      "epoch": 8.728877408244538,
      "grad_norm": 0.8824866414070129,
      "learning_rate": 7.655640555974572e-05,
      "loss": 3.1794,
      "step": 81100
    },
    {
      "epoch": 8.734258960284146,
      "grad_norm": 0.9138555526733398,
      "learning_rate": 7.623316452968428e-05,
      "loss": 3.1763,
      "step": 81150
    },
    {
      "epoch": 8.739640512323755,
      "grad_norm": 0.8968687653541565,
      "learning_rate": 7.590992349962288e-05,
      "loss": 3.1867,
      "step": 81200
    },
    {
      "epoch": 8.745022064363363,
      "grad_norm": 0.8379364609718323,
      "learning_rate": 7.558668246956147e-05,
      "loss": 3.1788,
      "step": 81250
    },
    {
      "epoch": 8.75040361640297,
      "grad_norm": 0.8813756108283997,
      "learning_rate": 7.526344143950005e-05,
      "loss": 3.1934,
      "step": 81300
    },
    {
      "epoch": 8.755785168442578,
      "grad_norm": 0.8551928400993347,
      "learning_rate": 7.494020040943862e-05,
      "loss": 3.1963,
      "step": 81350
    },
    {
      "epoch": 8.761166720482187,
      "grad_norm": 0.9404810667037964,
      "learning_rate": 7.461695937937722e-05,
      "loss": 3.1712,
      "step": 81400
    },
    {
      "epoch": 8.766548272521796,
      "grad_norm": 0.9553894400596619,
      "learning_rate": 7.42937183493158e-05,
      "loss": 3.1956,
      "step": 81450
    },
    {
      "epoch": 8.771929824561404,
      "grad_norm": 0.8532766103744507,
      "learning_rate": 7.397047731925439e-05,
      "loss": 3.1838,
      "step": 81500
    },
    {
      "epoch": 8.777311376601011,
      "grad_norm": 0.8783697485923767,
      "learning_rate": 7.364723628919297e-05,
      "loss": 3.1803,
      "step": 81550
    },
    {
      "epoch": 8.78269292864062,
      "grad_norm": 0.8748445510864258,
      "learning_rate": 7.332399525913155e-05,
      "loss": 3.1718,
      "step": 81600
    },
    {
      "epoch": 8.788074480680228,
      "grad_norm": 0.8972441554069519,
      "learning_rate": 7.300075422907013e-05,
      "loss": 3.1777,
      "step": 81650
    },
    {
      "epoch": 8.793456032719837,
      "grad_norm": 0.8829938173294067,
      "learning_rate": 7.267751319900872e-05,
      "loss": 3.1816,
      "step": 81700
    },
    {
      "epoch": 8.798837584759445,
      "grad_norm": 0.8357025980949402,
      "learning_rate": 7.235427216894731e-05,
      "loss": 3.1809,
      "step": 81750
    },
    {
      "epoch": 8.804219136799054,
      "grad_norm": 0.8655660152435303,
      "learning_rate": 7.203103113888589e-05,
      "loss": 3.1801,
      "step": 81800
    },
    {
      "epoch": 8.80960068883866,
      "grad_norm": 0.8427842855453491,
      "learning_rate": 7.17142549294257e-05,
      "loss": 3.1963,
      "step": 81850
    },
    {
      "epoch": 8.814982240878269,
      "grad_norm": 0.9213821887969971,
      "learning_rate": 7.139101389936428e-05,
      "loss": 3.1815,
      "step": 81900
    },
    {
      "epoch": 8.820363792917878,
      "grad_norm": 0.8830085396766663,
      "learning_rate": 7.10742376899041e-05,
      "loss": 3.1861,
      "step": 81950
    },
    {
      "epoch": 8.825745344957486,
      "grad_norm": 0.8858305811882019,
      "learning_rate": 7.075099665984268e-05,
      "loss": 3.1893,
      "step": 82000
    },
    {
      "epoch": 8.825745344957486,
      "eval_accuracy": 0.3927499839465358,
      "eval_loss": 3.313309669494629,
      "eval_runtime": 184.9366,
      "eval_samples_per_second": 97.39,
      "eval_steps_per_second": 6.089,
      "step": 82000
    },
    {
      "epoch": 8.831126896997095,
      "grad_norm": 0.898823082447052,
      "learning_rate": 7.042775562978126e-05,
      "loss": 3.1939,
      "step": 82050
    },
    {
      "epoch": 8.836508449036701,
      "grad_norm": 0.8878447413444519,
      "learning_rate": 7.010451459971986e-05,
      "loss": 3.1949,
      "step": 82100
    },
    {
      "epoch": 8.84189000107631,
      "grad_norm": 0.853378415107727,
      "learning_rate": 6.978127356965844e-05,
      "loss": 3.1804,
      "step": 82150
    },
    {
      "epoch": 8.847271553115919,
      "grad_norm": 0.8771511316299438,
      "learning_rate": 6.945803253959702e-05,
      "loss": 3.1889,
      "step": 82200
    },
    {
      "epoch": 8.852653105155527,
      "grad_norm": 0.8539415597915649,
      "learning_rate": 6.91347915095356e-05,
      "loss": 3.1983,
      "step": 82250
    },
    {
      "epoch": 8.858034657195136,
      "grad_norm": 0.8989981412887573,
      "learning_rate": 6.881155047947418e-05,
      "loss": 3.1973,
      "step": 82300
    },
    {
      "epoch": 8.863416209234742,
      "grad_norm": 0.8486732840538025,
      "learning_rate": 6.848830944941278e-05,
      "loss": 3.1764,
      "step": 82350
    },
    {
      "epoch": 8.868797761274351,
      "grad_norm": 0.862515389919281,
      "learning_rate": 6.816506841935136e-05,
      "loss": 3.1813,
      "step": 82400
    },
    {
      "epoch": 8.87417931331396,
      "grad_norm": 0.8676425814628601,
      "learning_rate": 6.784182738928994e-05,
      "loss": 3.1833,
      "step": 82450
    },
    {
      "epoch": 8.879560865353568,
      "grad_norm": 0.8382507562637329,
      "learning_rate": 6.751858635922853e-05,
      "loss": 3.1977,
      "step": 82500
    },
    {
      "epoch": 8.884942417393177,
      "grad_norm": 0.9167499542236328,
      "learning_rate": 6.719534532916711e-05,
      "loss": 3.2017,
      "step": 82550
    },
    {
      "epoch": 8.890323969432785,
      "grad_norm": 0.9321362972259521,
      "learning_rate": 6.68721042991057e-05,
      "loss": 3.1856,
      "step": 82600
    },
    {
      "epoch": 8.895705521472392,
      "grad_norm": 0.8884369134902954,
      "learning_rate": 6.654886326904428e-05,
      "loss": 3.166,
      "step": 82650
    },
    {
      "epoch": 8.901087073512,
      "grad_norm": 0.8399180769920349,
      "learning_rate": 6.622562223898286e-05,
      "loss": 3.1608,
      "step": 82700
    },
    {
      "epoch": 8.906468625551609,
      "grad_norm": 0.9197646379470825,
      "learning_rate": 6.590238120892145e-05,
      "loss": 3.19,
      "step": 82750
    },
    {
      "epoch": 8.911850177591218,
      "grad_norm": 0.8913666605949402,
      "learning_rate": 6.557914017886003e-05,
      "loss": 3.1812,
      "step": 82800
    },
    {
      "epoch": 8.917231729630826,
      "grad_norm": 0.9225286841392517,
      "learning_rate": 6.525589914879861e-05,
      "loss": 3.1826,
      "step": 82850
    },
    {
      "epoch": 8.922613281670433,
      "grad_norm": 0.8732419610023499,
      "learning_rate": 6.49326581187372e-05,
      "loss": 3.1804,
      "step": 82900
    },
    {
      "epoch": 8.927994833710041,
      "grad_norm": 0.9099151492118835,
      "learning_rate": 6.460941708867578e-05,
      "loss": 3.1793,
      "step": 82950
    },
    {
      "epoch": 8.93337638574965,
      "grad_norm": 0.8944827318191528,
      "learning_rate": 6.428617605861437e-05,
      "loss": 3.197,
      "step": 83000
    },
    {
      "epoch": 8.93337638574965,
      "eval_accuracy": 0.39288395295721107,
      "eval_loss": 3.311122417449951,
      "eval_runtime": 185.1062,
      "eval_samples_per_second": 97.301,
      "eval_steps_per_second": 6.083,
      "step": 83000
    },
    {
      "epoch": 8.938757937789259,
      "grad_norm": 0.9626566171646118,
      "learning_rate": 6.396293502855295e-05,
      "loss": 3.1786,
      "step": 83050
    },
    {
      "epoch": 8.944139489828867,
      "grad_norm": 0.8564572334289551,
      "learning_rate": 6.363969399849153e-05,
      "loss": 3.1572,
      "step": 83100
    },
    {
      "epoch": 8.949521041868476,
      "grad_norm": 0.8902326822280884,
      "learning_rate": 6.331645296843011e-05,
      "loss": 3.1862,
      "step": 83150
    },
    {
      "epoch": 8.954902593908082,
      "grad_norm": 0.8501969575881958,
      "learning_rate": 6.29932119383687e-05,
      "loss": 3.1809,
      "step": 83200
    },
    {
      "epoch": 8.960284145947691,
      "grad_norm": 0.8566688299179077,
      "learning_rate": 6.266997090830729e-05,
      "loss": 3.1817,
      "step": 83250
    },
    {
      "epoch": 8.9656656979873,
      "grad_norm": 0.8682833909988403,
      "learning_rate": 6.234672987824587e-05,
      "loss": 3.1946,
      "step": 83300
    },
    {
      "epoch": 8.971047250026908,
      "grad_norm": 0.9486375451087952,
      "learning_rate": 6.202348884818447e-05,
      "loss": 3.1855,
      "step": 83350
    },
    {
      "epoch": 8.976428802066517,
      "grad_norm": 0.8489530086517334,
      "learning_rate": 6.170024781812305e-05,
      "loss": 3.181,
      "step": 83400
    },
    {
      "epoch": 8.981810354106123,
      "grad_norm": 0.8610159158706665,
      "learning_rate": 6.137700678806163e-05,
      "loss": 3.1787,
      "step": 83450
    },
    {
      "epoch": 8.987191906145732,
      "grad_norm": 0.926662802696228,
      "learning_rate": 6.105376575800021e-05,
      "loss": 3.1692,
      "step": 83500
    },
    {
      "epoch": 8.99257345818534,
      "grad_norm": 0.9315507411956787,
      "learning_rate": 6.073052472793879e-05,
      "loss": 3.1682,
      "step": 83550
    },
    {
      "epoch": 8.997955010224949,
      "grad_norm": 0.9029734134674072,
      "learning_rate": 6.0407283697877384e-05,
      "loss": 3.1876,
      "step": 83600
    },
    {
      "epoch": 9.003336562264558,
      "grad_norm": 0.9033173322677612,
      "learning_rate": 6.0084042667815966e-05,
      "loss": 3.1469,
      "step": 83650
    },
    {
      "epoch": 9.008718114304166,
      "grad_norm": 0.8773213028907776,
      "learning_rate": 5.976080163775455e-05,
      "loss": 3.1173,
      "step": 83700
    },
    {
      "epoch": 9.014099666343773,
      "grad_norm": 0.9068469405174255,
      "learning_rate": 5.9437560607693135e-05,
      "loss": 3.1375,
      "step": 83750
    },
    {
      "epoch": 9.019481218383381,
      "grad_norm": 0.8826611042022705,
      "learning_rate": 5.9114319577631716e-05,
      "loss": 3.0867,
      "step": 83800
    },
    {
      "epoch": 9.02486277042299,
      "grad_norm": 0.8884468674659729,
      "learning_rate": 5.8791078547570304e-05,
      "loss": 3.1317,
      "step": 83850
    },
    {
      "epoch": 9.030244322462599,
      "grad_norm": 0.8935588002204895,
      "learning_rate": 5.8467837517508885e-05,
      "loss": 3.1211,
      "step": 83900
    },
    {
      "epoch": 9.035625874502207,
      "grad_norm": 0.8488547205924988,
      "learning_rate": 5.8144596487447466e-05,
      "loss": 3.1253,
      "step": 83950
    },
    {
      "epoch": 9.041007426541814,
      "grad_norm": 0.9119958877563477,
      "learning_rate": 5.7821355457386054e-05,
      "loss": 3.1188,
      "step": 84000
    },
    {
      "epoch": 9.041007426541814,
      "eval_accuracy": 0.3928714578751124,
      "eval_loss": 3.3146016597747803,
      "eval_runtime": 184.5536,
      "eval_samples_per_second": 97.592,
      "eval_steps_per_second": 6.101,
      "step": 84000
    },
    {
      "epoch": 9.046388978581422,
      "grad_norm": 0.8722310662269592,
      "learning_rate": 5.7498114427324635e-05,
      "loss": 3.1142,
      "step": 84050
    },
    {
      "epoch": 9.051770530621031,
      "grad_norm": 0.8786686062812805,
      "learning_rate": 5.717487339726322e-05,
      "loss": 3.1196,
      "step": 84100
    },
    {
      "epoch": 9.05715208266064,
      "grad_norm": 0.873644232749939,
      "learning_rate": 5.6851632367201804e-05,
      "loss": 3.1298,
      "step": 84150
    },
    {
      "epoch": 9.062533634700248,
      "grad_norm": 0.8933513760566711,
      "learning_rate": 5.6528391337140385e-05,
      "loss": 3.1089,
      "step": 84200
    },
    {
      "epoch": 9.067915186739857,
      "grad_norm": 0.8709551095962524,
      "learning_rate": 5.620515030707897e-05,
      "loss": 3.106,
      "step": 84250
    },
    {
      "epoch": 9.073296738779463,
      "grad_norm": 0.8717238306999207,
      "learning_rate": 5.5881909277017554e-05,
      "loss": 3.1345,
      "step": 84300
    },
    {
      "epoch": 9.078678290819072,
      "grad_norm": 0.858155369758606,
      "learning_rate": 5.555866824695614e-05,
      "loss": 3.1032,
      "step": 84350
    },
    {
      "epoch": 9.08405984285868,
      "grad_norm": 0.8939986824989319,
      "learning_rate": 5.523542721689472e-05,
      "loss": 3.1245,
      "step": 84400
    },
    {
      "epoch": 9.089441394898289,
      "grad_norm": 0.9245389103889465,
      "learning_rate": 5.4912186186833304e-05,
      "loss": 3.1343,
      "step": 84450
    },
    {
      "epoch": 9.094822946937898,
      "grad_norm": 0.9213153719902039,
      "learning_rate": 5.45889451567719e-05,
      "loss": 3.1191,
      "step": 84500
    },
    {
      "epoch": 9.100204498977504,
      "grad_norm": 0.8970081210136414,
      "learning_rate": 5.426570412671048e-05,
      "loss": 3.1257,
      "step": 84550
    },
    {
      "epoch": 9.105586051017113,
      "grad_norm": 0.9398916959762573,
      "learning_rate": 5.394246309664907e-05,
      "loss": 3.1261,
      "step": 84600
    },
    {
      "epoch": 9.110967603056721,
      "grad_norm": 0.8600006103515625,
      "learning_rate": 5.361922206658765e-05,
      "loss": 3.1337,
      "step": 84650
    },
    {
      "epoch": 9.11634915509633,
      "grad_norm": 0.926282525062561,
      "learning_rate": 5.329598103652623e-05,
      "loss": 3.1305,
      "step": 84700
    },
    {
      "epoch": 9.121730707135939,
      "grad_norm": 0.8645695447921753,
      "learning_rate": 5.297274000646482e-05,
      "loss": 3.1233,
      "step": 84750
    },
    {
      "epoch": 9.127112259175545,
      "grad_norm": 0.8621381521224976,
      "learning_rate": 5.26494989764034e-05,
      "loss": 3.121,
      "step": 84800
    },
    {
      "epoch": 9.132493811215154,
      "grad_norm": 0.8733987212181091,
      "learning_rate": 5.232625794634199e-05,
      "loss": 3.1233,
      "step": 84850
    },
    {
      "epoch": 9.137875363254762,
      "grad_norm": 0.940946102142334,
      "learning_rate": 5.200301691628057e-05,
      "loss": 3.1294,
      "step": 84900
    },
    {
      "epoch": 9.143256915294371,
      "grad_norm": 0.8964011073112488,
      "learning_rate": 5.167977588621915e-05,
      "loss": 3.1282,
      "step": 84950
    },
    {
      "epoch": 9.14863846733398,
      "grad_norm": 0.9262023568153381,
      "learning_rate": 5.135653485615774e-05,
      "loss": 3.1232,
      "step": 85000
    },
    {
      "epoch": 9.14863846733398,
      "eval_accuracy": 0.3931999155549756,
      "eval_loss": 3.31473970413208,
      "eval_runtime": 184.9896,
      "eval_samples_per_second": 97.362,
      "eval_steps_per_second": 6.087,
      "step": 85000
    },
    {
      "epoch": 9.154020019373588,
      "grad_norm": 0.860278844833374,
      "learning_rate": 5.103329382609632e-05,
      "loss": 3.1191,
      "step": 85050
    },
    {
      "epoch": 9.159401571413195,
      "grad_norm": 0.9099411368370056,
      "learning_rate": 5.0710052796034906e-05,
      "loss": 3.1229,
      "step": 85100
    },
    {
      "epoch": 9.164783123452803,
      "grad_norm": 0.8791196942329407,
      "learning_rate": 5.038681176597349e-05,
      "loss": 3.0978,
      "step": 85150
    },
    {
      "epoch": 9.170164675492412,
      "grad_norm": 1.0391086339950562,
      "learning_rate": 5.006357073591207e-05,
      "loss": 3.1267,
      "step": 85200
    },
    {
      "epoch": 9.17554622753202,
      "grad_norm": 0.8792301416397095,
      "learning_rate": 4.974032970585066e-05,
      "loss": 3.1279,
      "step": 85250
    },
    {
      "epoch": 9.180927779571629,
      "grad_norm": 0.8521240949630737,
      "learning_rate": 4.9417088675789244e-05,
      "loss": 3.118,
      "step": 85300
    },
    {
      "epoch": 9.186309331611236,
      "grad_norm": 0.912763237953186,
      "learning_rate": 4.909384764572783e-05,
      "loss": 3.121,
      "step": 85350
    },
    {
      "epoch": 9.191690883650844,
      "grad_norm": 0.8790477514266968,
      "learning_rate": 4.877060661566641e-05,
      "loss": 3.129,
      "step": 85400
    },
    {
      "epoch": 9.197072435690453,
      "grad_norm": 0.8999446630477905,
      "learning_rate": 4.8447365585604994e-05,
      "loss": 3.1379,
      "step": 85450
    },
    {
      "epoch": 9.202453987730062,
      "grad_norm": 0.8669745326042175,
      "learning_rate": 4.812412455554358e-05,
      "loss": 3.1175,
      "step": 85500
    },
    {
      "epoch": 9.20783553976967,
      "grad_norm": 0.9326590895652771,
      "learning_rate": 4.780088352548216e-05,
      "loss": 3.1152,
      "step": 85550
    },
    {
      "epoch": 9.213217091809279,
      "grad_norm": 0.8711249828338623,
      "learning_rate": 4.747764249542075e-05,
      "loss": 3.1119,
      "step": 85600
    },
    {
      "epoch": 9.218598643848885,
      "grad_norm": 0.9170059561729431,
      "learning_rate": 4.715440146535933e-05,
      "loss": 3.1528,
      "step": 85650
    },
    {
      "epoch": 9.223980195888494,
      "grad_norm": 0.8670192360877991,
      "learning_rate": 4.683116043529791e-05,
      "loss": 3.1398,
      "step": 85700
    },
    {
      "epoch": 9.229361747928102,
      "grad_norm": 0.9062800407409668,
      "learning_rate": 4.65079194052365e-05,
      "loss": 3.1122,
      "step": 85750
    },
    {
      "epoch": 9.234743299967711,
      "grad_norm": 0.8952690362930298,
      "learning_rate": 4.618467837517508e-05,
      "loss": 3.1347,
      "step": 85800
    },
    {
      "epoch": 9.24012485200732,
      "grad_norm": 0.9159784317016602,
      "learning_rate": 4.586143734511367e-05,
      "loss": 3.1263,
      "step": 85850
    },
    {
      "epoch": 9.245506404046926,
      "grad_norm": 0.9201834797859192,
      "learning_rate": 4.553819631505225e-05,
      "loss": 3.1292,
      "step": 85900
    },
    {
      "epoch": 9.250887956086535,
      "grad_norm": 0.8721088767051697,
      "learning_rate": 4.521495528499083e-05,
      "loss": 3.126,
      "step": 85950
    },
    {
      "epoch": 9.256269508126143,
      "grad_norm": 0.9054554104804993,
      "learning_rate": 4.4898179075530645e-05,
      "loss": 3.1126,
      "step": 86000
    },
    {
      "epoch": 9.256269508126143,
      "eval_accuracy": 0.3934344971397671,
      "eval_loss": 3.3119852542877197,
      "eval_runtime": 184.8448,
      "eval_samples_per_second": 97.438,
      "eval_steps_per_second": 6.092,
      "step": 86000
    },
    {
      "epoch": 9.261651060165752,
      "grad_norm": 0.8648436665534973,
      "learning_rate": 4.457493804546923e-05,
      "loss": 3.1399,
      "step": 86050
    },
    {
      "epoch": 9.26703261220536,
      "grad_norm": 0.8899005651473999,
      "learning_rate": 4.4251697015407814e-05,
      "loss": 3.1283,
      "step": 86100
    },
    {
      "epoch": 9.272414164244967,
      "grad_norm": 0.8872169852256775,
      "learning_rate": 4.392845598534641e-05,
      "loss": 3.127,
      "step": 86150
    },
    {
      "epoch": 9.277795716284576,
      "grad_norm": 0.8724729418754578,
      "learning_rate": 4.360521495528499e-05,
      "loss": 3.126,
      "step": 86200
    },
    {
      "epoch": 9.283177268324184,
      "grad_norm": 0.8950654864311218,
      "learning_rate": 4.3281973925223564e-05,
      "loss": 3.1202,
      "step": 86250
    },
    {
      "epoch": 9.288558820363793,
      "grad_norm": 0.9006926417350769,
      "learning_rate": 4.295873289516216e-05,
      "loss": 3.1234,
      "step": 86300
    },
    {
      "epoch": 9.293940372403402,
      "grad_norm": 0.8863205313682556,
      "learning_rate": 4.263549186510074e-05,
      "loss": 3.112,
      "step": 86350
    },
    {
      "epoch": 9.29932192444301,
      "grad_norm": 0.8906213045120239,
      "learning_rate": 4.231225083503933e-05,
      "loss": 3.1287,
      "step": 86400
    },
    {
      "epoch": 9.304703476482617,
      "grad_norm": 0.8871445655822754,
      "learning_rate": 4.198900980497791e-05,
      "loss": 3.1298,
      "step": 86450
    },
    {
      "epoch": 9.310085028522225,
      "grad_norm": 0.9487957954406738,
      "learning_rate": 4.166576877491649e-05,
      "loss": 3.1195,
      "step": 86500
    },
    {
      "epoch": 9.315466580561834,
      "grad_norm": 0.8825381398200989,
      "learning_rate": 4.134252774485508e-05,
      "loss": 3.1436,
      "step": 86550
    },
    {
      "epoch": 9.320848132601443,
      "grad_norm": 0.8810291886329651,
      "learning_rate": 4.101928671479366e-05,
      "loss": 3.1323,
      "step": 86600
    },
    {
      "epoch": 9.326229684641051,
      "grad_norm": 0.9148359298706055,
      "learning_rate": 4.069604568473225e-05,
      "loss": 3.1344,
      "step": 86650
    },
    {
      "epoch": 9.331611236680658,
      "grad_norm": 0.8992313146591187,
      "learning_rate": 4.037280465467083e-05,
      "loss": 3.1286,
      "step": 86700
    },
    {
      "epoch": 9.336992788720266,
      "grad_norm": 0.9089197516441345,
      "learning_rate": 4.004956362460941e-05,
      "loss": 3.1332,
      "step": 86750
    },
    {
      "epoch": 9.342374340759875,
      "grad_norm": 0.868971586227417,
      "learning_rate": 3.9726322594548e-05,
      "loss": 3.1276,
      "step": 86800
    },
    {
      "epoch": 9.347755892799483,
      "grad_norm": 0.921055257320404,
      "learning_rate": 3.940308156448658e-05,
      "loss": 3.1217,
      "step": 86850
    },
    {
      "epoch": 9.353137444839092,
      "grad_norm": 0.959754228591919,
      "learning_rate": 3.9079840534425166e-05,
      "loss": 3.1177,
      "step": 86900
    },
    {
      "epoch": 9.3585189968787,
      "grad_norm": 0.892400324344635,
      "learning_rate": 3.875659950436375e-05,
      "loss": 3.1181,
      "step": 86950
    },
    {
      "epoch": 9.363900548918307,
      "grad_norm": 0.9151771068572998,
      "learning_rate": 3.843335847430233e-05,
      "loss": 3.1101,
      "step": 87000
    },
    {
      "epoch": 9.363900548918307,
      "eval_accuracy": 0.3936111667353534,
      "eval_loss": 3.3083672523498535,
      "eval_runtime": 185.0906,
      "eval_samples_per_second": 97.309,
      "eval_steps_per_second": 6.084,
      "step": 87000
    },
    {
      "epoch": 9.369282100957916,
      "grad_norm": 0.8925448656082153,
      "learning_rate": 3.811011744424092e-05,
      "loss": 3.1359,
      "step": 87050
    },
    {
      "epoch": 9.374663652997524,
      "grad_norm": 0.8919752240180969,
      "learning_rate": 3.7786876414179504e-05,
      "loss": 3.1191,
      "step": 87100
    },
    {
      "epoch": 9.380045205037133,
      "grad_norm": 0.9180679321289062,
      "learning_rate": 3.7463635384118085e-05,
      "loss": 3.1496,
      "step": 87150
    },
    {
      "epoch": 9.385426757076742,
      "grad_norm": 0.8926573991775513,
      "learning_rate": 3.714039435405667e-05,
      "loss": 3.1319,
      "step": 87200
    },
    {
      "epoch": 9.390808309116348,
      "grad_norm": 0.9110729098320007,
      "learning_rate": 3.6817153323995254e-05,
      "loss": 3.1211,
      "step": 87250
    },
    {
      "epoch": 9.396189861155957,
      "grad_norm": 0.9294565320014954,
      "learning_rate": 3.649391229393384e-05,
      "loss": 3.1246,
      "step": 87300
    },
    {
      "epoch": 9.401571413195565,
      "grad_norm": 0.88547283411026,
      "learning_rate": 3.617067126387242e-05,
      "loss": 3.137,
      "step": 87350
    },
    {
      "epoch": 9.406952965235174,
      "grad_norm": 0.8936682343482971,
      "learning_rate": 3.5847430233811004e-05,
      "loss": 3.1245,
      "step": 87400
    },
    {
      "epoch": 9.412334517274783,
      "grad_norm": 0.9498528838157654,
      "learning_rate": 3.552418920374959e-05,
      "loss": 3.1319,
      "step": 87450
    },
    {
      "epoch": 9.417716069314391,
      "grad_norm": 0.949592113494873,
      "learning_rate": 3.520094817368818e-05,
      "loss": 3.1445,
      "step": 87500
    },
    {
      "epoch": 9.423097621353998,
      "grad_norm": 0.8684675097465515,
      "learning_rate": 3.487770714362676e-05,
      "loss": 3.1472,
      "step": 87550
    },
    {
      "epoch": 9.428479173393606,
      "grad_norm": 0.950598418712616,
      "learning_rate": 3.455446611356535e-05,
      "loss": 3.1463,
      "step": 87600
    },
    {
      "epoch": 9.433860725433215,
      "grad_norm": 0.9288617968559265,
      "learning_rate": 3.423122508350393e-05,
      "loss": 3.1354,
      "step": 87650
    },
    {
      "epoch": 9.439242277472824,
      "grad_norm": 0.9261043071746826,
      "learning_rate": 3.390798405344251e-05,
      "loss": 3.1414,
      "step": 87700
    },
    {
      "epoch": 9.444623829512432,
      "grad_norm": 0.8761609196662903,
      "learning_rate": 3.35847430233811e-05,
      "loss": 3.1457,
      "step": 87750
    },
    {
      "epoch": 9.450005381552039,
      "grad_norm": 0.9005210995674133,
      "learning_rate": 3.326150199331968e-05,
      "loss": 3.1285,
      "step": 87800
    },
    {
      "epoch": 9.455386933591647,
      "grad_norm": 0.9317635297775269,
      "learning_rate": 3.293826096325827e-05,
      "loss": 3.1293,
      "step": 87850
    },
    {
      "epoch": 9.460768485631256,
      "grad_norm": 0.886011004447937,
      "learning_rate": 3.261501993319685e-05,
      "loss": 3.1344,
      "step": 87900
    },
    {
      "epoch": 9.466150037670864,
      "grad_norm": 0.9412506818771362,
      "learning_rate": 3.229177890313544e-05,
      "loss": 3.136,
      "step": 87950
    },
    {
      "epoch": 9.471531589710473,
      "grad_norm": 0.9001096487045288,
      "learning_rate": 3.197500269367525e-05,
      "loss": 3.1343,
      "step": 88000
    },
    {
      "epoch": 9.471531589710473,
      "eval_accuracy": 0.39403382646895174,
      "eval_loss": 3.306755781173706,
      "eval_runtime": 184.7543,
      "eval_samples_per_second": 97.486,
      "eval_steps_per_second": 6.095,
      "step": 88000
    },
    {
      "epoch": 9.476913141750082,
      "grad_norm": 0.9205313920974731,
      "learning_rate": 3.165176166361383e-05,
      "loss": 3.1289,
      "step": 88050
    },
    {
      "epoch": 9.482294693789688,
      "grad_norm": 0.9257974624633789,
      "learning_rate": 3.132852063355242e-05,
      "loss": 3.1247,
      "step": 88100
    },
    {
      "epoch": 9.487676245829297,
      "grad_norm": 0.9012815952301025,
      "learning_rate": 3.1005279603491e-05,
      "loss": 3.1413,
      "step": 88150
    },
    {
      "epoch": 9.493057797868905,
      "grad_norm": 0.9336722493171692,
      "learning_rate": 3.068203857342958e-05,
      "loss": 3.1185,
      "step": 88200
    },
    {
      "epoch": 9.498439349908514,
      "grad_norm": 0.9232977628707886,
      "learning_rate": 3.035879754336817e-05,
      "loss": 3.1124,
      "step": 88250
    },
    {
      "epoch": 9.503820901948123,
      "grad_norm": 0.8725977540016174,
      "learning_rate": 3.0035556513306754e-05,
      "loss": 3.1293,
      "step": 88300
    },
    {
      "epoch": 9.50920245398773,
      "grad_norm": 0.9048357009887695,
      "learning_rate": 2.971231548324534e-05,
      "loss": 3.1483,
      "step": 88350
    },
    {
      "epoch": 9.514584006027338,
      "grad_norm": 0.9110617637634277,
      "learning_rate": 2.938907445318392e-05,
      "loss": 3.1158,
      "step": 88400
    },
    {
      "epoch": 9.519965558066946,
      "grad_norm": 0.8577834963798523,
      "learning_rate": 2.9065833423122504e-05,
      "loss": 3.1259,
      "step": 88450
    },
    {
      "epoch": 9.525347110106555,
      "grad_norm": 0.8744284510612488,
      "learning_rate": 2.874259239306109e-05,
      "loss": 3.1197,
      "step": 88500
    },
    {
      "epoch": 9.530728662146164,
      "grad_norm": 0.8887792229652405,
      "learning_rate": 2.8419351362999673e-05,
      "loss": 3.1283,
      "step": 88550
    },
    {
      "epoch": 9.536110214185772,
      "grad_norm": 0.9559165835380554,
      "learning_rate": 2.809611033293826e-05,
      "loss": 3.1234,
      "step": 88600
    },
    {
      "epoch": 9.541491766225379,
      "grad_norm": 0.8760658502578735,
      "learning_rate": 2.7772869302876845e-05,
      "loss": 3.118,
      "step": 88650
    },
    {
      "epoch": 9.546873318264987,
      "grad_norm": 0.885046660900116,
      "learning_rate": 2.7449628272815427e-05,
      "loss": 3.1529,
      "step": 88700
    },
    {
      "epoch": 9.552254870304596,
      "grad_norm": 0.8885508179664612,
      "learning_rate": 2.712638724275401e-05,
      "loss": 3.1341,
      "step": 88750
    },
    {
      "epoch": 9.557636422344205,
      "grad_norm": 0.90824294090271,
      "learning_rate": 2.6803146212692596e-05,
      "loss": 3.1306,
      "step": 88800
    },
    {
      "epoch": 9.563017974383813,
      "grad_norm": 0.9239144921302795,
      "learning_rate": 2.647990518263118e-05,
      "loss": 3.1522,
      "step": 88850
    },
    {
      "epoch": 9.56839952642342,
      "grad_norm": 0.927269458770752,
      "learning_rate": 2.6156664152569765e-05,
      "loss": 3.1294,
      "step": 88900
    },
    {
      "epoch": 9.573781078463028,
      "grad_norm": 0.8974427580833435,
      "learning_rate": 2.5833423122508346e-05,
      "loss": 3.1244,
      "step": 88950
    },
    {
      "epoch": 9.579162630502637,
      "grad_norm": 0.844535231590271,
      "learning_rate": 2.5510182092446934e-05,
      "loss": 3.1383,
      "step": 89000
    },
    {
      "epoch": 9.579162630502637,
      "eval_accuracy": 0.39431393361373746,
      "eval_loss": 3.303724765777588,
      "eval_runtime": 184.8828,
      "eval_samples_per_second": 97.418,
      "eval_steps_per_second": 6.09,
      "step": 89000
    },
    {
      "epoch": 9.584544182542245,
      "grad_norm": 0.8630052804946899,
      "learning_rate": 2.5186941062385518e-05,
      "loss": 3.1265,
      "step": 89050
    },
    {
      "epoch": 9.589925734581854,
      "grad_norm": 0.8742738366127014,
      "learning_rate": 2.4863700032324103e-05,
      "loss": 3.144,
      "step": 89100
    },
    {
      "epoch": 9.59530728662146,
      "grad_norm": 0.9079467058181763,
      "learning_rate": 2.4540459002262687e-05,
      "loss": 3.1231,
      "step": 89150
    },
    {
      "epoch": 9.60068883866107,
      "grad_norm": 0.8864733576774597,
      "learning_rate": 2.4217217972201268e-05,
      "loss": 3.1178,
      "step": 89200
    },
    {
      "epoch": 9.606070390700678,
      "grad_norm": 0.9138379096984863,
      "learning_rate": 2.3893976942139853e-05,
      "loss": 3.1156,
      "step": 89250
    },
    {
      "epoch": 9.611451942740286,
      "grad_norm": 0.9160833954811096,
      "learning_rate": 2.3570735912078437e-05,
      "loss": 3.1396,
      "step": 89300
    },
    {
      "epoch": 9.616833494779895,
      "grad_norm": 0.9601799249649048,
      "learning_rate": 2.324749488201702e-05,
      "loss": 3.1191,
      "step": 89350
    },
    {
      "epoch": 9.622215046819504,
      "grad_norm": 0.9554757475852966,
      "learning_rate": 2.292425385195561e-05,
      "loss": 3.1543,
      "step": 89400
    },
    {
      "epoch": 9.62759659885911,
      "grad_norm": 0.9189677238464355,
      "learning_rate": 2.260101282189419e-05,
      "loss": 3.1164,
      "step": 89450
    },
    {
      "epoch": 9.632978150898719,
      "grad_norm": 0.888224184513092,
      "learning_rate": 2.2277771791832775e-05,
      "loss": 3.1328,
      "step": 89500
    },
    {
      "epoch": 9.638359702938327,
      "grad_norm": 0.8766180872917175,
      "learning_rate": 2.195453076177136e-05,
      "loss": 3.1125,
      "step": 89550
    },
    {
      "epoch": 9.643741254977936,
      "grad_norm": 0.9146680235862732,
      "learning_rate": 2.1631289731709944e-05,
      "loss": 3.121,
      "step": 89600
    },
    {
      "epoch": 9.649122807017545,
      "grad_norm": 0.9508671760559082,
      "learning_rate": 2.130804870164853e-05,
      "loss": 3.1331,
      "step": 89650
    },
    {
      "epoch": 9.654504359057151,
      "grad_norm": 0.8772718906402588,
      "learning_rate": 2.098480767158711e-05,
      "loss": 3.1178,
      "step": 89700
    },
    {
      "epoch": 9.65988591109676,
      "grad_norm": 0.9173164963722229,
      "learning_rate": 2.0661566641525694e-05,
      "loss": 3.1261,
      "step": 89750
    },
    {
      "epoch": 9.665267463136368,
      "grad_norm": 0.8904097676277161,
      "learning_rate": 2.0338325611464282e-05,
      "loss": 3.1378,
      "step": 89800
    },
    {
      "epoch": 9.670649015175977,
      "grad_norm": 0.9122865200042725,
      "learning_rate": 2.0015084581402867e-05,
      "loss": 3.1482,
      "step": 89850
    },
    {
      "epoch": 9.676030567215586,
      "grad_norm": 0.9031899571418762,
      "learning_rate": 1.969184355134145e-05,
      "loss": 3.1317,
      "step": 89900
    },
    {
      "epoch": 9.681412119255192,
      "grad_norm": 0.9516944885253906,
      "learning_rate": 1.9368602521280032e-05,
      "loss": 3.1399,
      "step": 89950
    },
    {
      "epoch": 9.6867936712948,
      "grad_norm": 0.8837878108024597,
      "learning_rate": 1.9045361491218617e-05,
      "loss": 3.1249,
      "step": 90000
    },
    {
      "epoch": 9.6867936712948,
      "eval_accuracy": 0.39447756486278607,
      "eval_loss": 3.3018171787261963,
      "eval_runtime": 185.0594,
      "eval_samples_per_second": 97.326,
      "eval_steps_per_second": 6.085,
      "step": 90000
    },
    {
      "epoch": 9.69217522333441,
      "grad_norm": 0.9284664988517761,
      "learning_rate": 1.872858528175843e-05,
      "loss": 3.1494,
      "step": 90050
    },
    {
      "epoch": 9.697556775374018,
      "grad_norm": 0.9753976464271545,
      "learning_rate": 1.8405344251697014e-05,
      "loss": 3.1438,
      "step": 90100
    },
    {
      "epoch": 9.702938327413626,
      "grad_norm": 0.9227730631828308,
      "learning_rate": 1.80821032216356e-05,
      "loss": 3.1289,
      "step": 90150
    },
    {
      "epoch": 9.708319879453235,
      "grad_norm": 0.8707805275917053,
      "learning_rate": 1.7758862191574183e-05,
      "loss": 3.1446,
      "step": 90200
    },
    {
      "epoch": 9.713701431492842,
      "grad_norm": 0.8503866791725159,
      "learning_rate": 1.7435621161512768e-05,
      "loss": 3.1292,
      "step": 90250
    },
    {
      "epoch": 9.71908298353245,
      "grad_norm": 0.8727654218673706,
      "learning_rate": 1.711238013145135e-05,
      "loss": 3.121,
      "step": 90300
    },
    {
      "epoch": 9.724464535572059,
      "grad_norm": 0.9083965420722961,
      "learning_rate": 1.6789139101389937e-05,
      "loss": 3.1278,
      "step": 90350
    },
    {
      "epoch": 9.729846087611667,
      "grad_norm": 0.865218997001648,
      "learning_rate": 1.6472362891929747e-05,
      "loss": 3.1158,
      "step": 90400
    },
    {
      "epoch": 9.735227639651276,
      "grad_norm": 0.9163516759872437,
      "learning_rate": 1.614912186186833e-05,
      "loss": 3.1345,
      "step": 90450
    },
    {
      "epoch": 9.740609191690883,
      "grad_norm": 0.8571822047233582,
      "learning_rate": 1.5825880831806916e-05,
      "loss": 3.1355,
      "step": 90500
    },
    {
      "epoch": 9.745990743730491,
      "grad_norm": 0.8619289398193359,
      "learning_rate": 1.55026398017455e-05,
      "loss": 3.113,
      "step": 90550
    },
    {
      "epoch": 9.7513722957701,
      "grad_norm": 0.8706497550010681,
      "learning_rate": 1.5179398771684085e-05,
      "loss": 3.1324,
      "step": 90600
    },
    {
      "epoch": 9.756753847809708,
      "grad_norm": 0.9034489989280701,
      "learning_rate": 1.485615774162267e-05,
      "loss": 3.1221,
      "step": 90650
    },
    {
      "epoch": 9.762135399849317,
      "grad_norm": 0.8666117191314697,
      "learning_rate": 1.4532916711561252e-05,
      "loss": 3.1205,
      "step": 90700
    },
    {
      "epoch": 9.767516951888926,
      "grad_norm": 0.9398617148399353,
      "learning_rate": 1.4209675681499837e-05,
      "loss": 3.1403,
      "step": 90750
    },
    {
      "epoch": 9.772898503928532,
      "grad_norm": 0.9415978789329529,
      "learning_rate": 1.3886434651438423e-05,
      "loss": 3.1329,
      "step": 90800
    },
    {
      "epoch": 9.77828005596814,
      "grad_norm": 0.8788616061210632,
      "learning_rate": 1.3563193621377006e-05,
      "loss": 3.1208,
      "step": 90850
    },
    {
      "epoch": 9.78366160800775,
      "grad_norm": 0.8646774888038635,
      "learning_rate": 1.323995259131559e-05,
      "loss": 3.1319,
      "step": 90900
    },
    {
      "epoch": 9.789043160047358,
      "grad_norm": 0.9247779846191406,
      "learning_rate": 1.2916711561254173e-05,
      "loss": 3.1319,
      "step": 90950
    },
    {
      "epoch": 9.794424712086967,
      "grad_norm": 0.9060809016227722,
      "learning_rate": 1.2593470531192759e-05,
      "loss": 3.1263,
      "step": 91000
    },
    {
      "epoch": 9.794424712086967,
      "eval_accuracy": 0.39467879001101847,
      "eval_loss": 3.299553394317627,
      "eval_runtime": 184.8969,
      "eval_samples_per_second": 97.411,
      "eval_steps_per_second": 6.09,
      "step": 91000
    },
    {
      "epoch": 9.799806264126573,
      "grad_norm": 0.8818645477294922,
      "learning_rate": 1.2270229501131344e-05,
      "loss": 3.1265,
      "step": 91050
    },
    {
      "epoch": 9.805187816166182,
      "grad_norm": 0.917174220085144,
      "learning_rate": 1.1946988471069926e-05,
      "loss": 3.1284,
      "step": 91100
    },
    {
      "epoch": 9.81056936820579,
      "grad_norm": 0.9325860738754272,
      "learning_rate": 1.162374744100851e-05,
      "loss": 3.1433,
      "step": 91150
    },
    {
      "epoch": 9.815950920245399,
      "grad_norm": 0.9602986574172974,
      "learning_rate": 1.1300506410947095e-05,
      "loss": 3.1445,
      "step": 91200
    },
    {
      "epoch": 9.821332472285007,
      "grad_norm": 0.9245926737785339,
      "learning_rate": 1.097726538088568e-05,
      "loss": 3.1266,
      "step": 91250
    },
    {
      "epoch": 9.826714024324616,
      "grad_norm": 0.9178900718688965,
      "learning_rate": 1.0654024350824264e-05,
      "loss": 3.1296,
      "step": 91300
    },
    {
      "epoch": 9.832095576364223,
      "grad_norm": 0.8924757838249207,
      "learning_rate": 1.0330783320762847e-05,
      "loss": 3.1316,
      "step": 91350
    },
    {
      "epoch": 9.837477128403831,
      "grad_norm": 0.8745411038398743,
      "learning_rate": 1.0007542290701433e-05,
      "loss": 3.1478,
      "step": 91400
    },
    {
      "epoch": 9.84285868044344,
      "grad_norm": 0.9549648761749268,
      "learning_rate": 9.684301260640016e-06,
      "loss": 3.1255,
      "step": 91450
    },
    {
      "epoch": 9.848240232483048,
      "grad_norm": 0.8879762291908264,
      "learning_rate": 9.3610602305786e-06,
      "loss": 3.1295,
      "step": 91500
    },
    {
      "epoch": 9.853621784522657,
      "grad_norm": 0.8747987747192383,
      "learning_rate": 9.037819200517185e-06,
      "loss": 3.1141,
      "step": 91550
    },
    {
      "epoch": 9.859003336562264,
      "grad_norm": 0.9374912977218628,
      "learning_rate": 8.71457817045577e-06,
      "loss": 3.133,
      "step": 91600
    },
    {
      "epoch": 9.864384888601872,
      "grad_norm": 0.8693592548370361,
      "learning_rate": 8.391337140394354e-06,
      "loss": 3.1176,
      "step": 91650
    },
    {
      "epoch": 9.869766440641481,
      "grad_norm": 0.9023551344871521,
      "learning_rate": 8.068096110332939e-06,
      "loss": 3.116,
      "step": 91700
    },
    {
      "epoch": 9.87514799268109,
      "grad_norm": 0.8818048238754272,
      "learning_rate": 7.744855080271521e-06,
      "loss": 3.134,
      "step": 91750
    },
    {
      "epoch": 9.880529544720698,
      "grad_norm": 0.9004733562469482,
      "learning_rate": 7.421614050210106e-06,
      "loss": 3.1342,
      "step": 91800
    },
    {
      "epoch": 9.885911096760307,
      "grad_norm": 0.9453480243682861,
      "learning_rate": 7.09837302014869e-06,
      "loss": 3.1202,
      "step": 91850
    },
    {
      "epoch": 9.891292648799913,
      "grad_norm": 0.8884382843971252,
      "learning_rate": 6.775131990087275e-06,
      "loss": 3.1268,
      "step": 91900
    },
    {
      "epoch": 9.896674200839522,
      "grad_norm": 0.9106020927429199,
      "learning_rate": 6.451890960025859e-06,
      "loss": 3.1243,
      "step": 91950
    },
    {
      "epoch": 9.90205575287913,
      "grad_norm": 0.9381679892539978,
      "learning_rate": 6.128649929964443e-06,
      "loss": 3.1164,
      "step": 92000
    },
    {
      "epoch": 9.90205575287913,
      "eval_accuracy": 0.39488544780364165,
      "eval_loss": 3.2985033988952637,
      "eval_runtime": 184.9441,
      "eval_samples_per_second": 97.386,
      "eval_steps_per_second": 6.088,
      "step": 92000
    },
    {
      "epoch": 9.907437304918739,
      "grad_norm": 0.9050652384757996,
      "learning_rate": 5.805408899903027e-06,
      "loss": 3.1365,
      "step": 92050
    },
    {
      "epoch": 9.912818856958348,
      "grad_norm": 0.924476146697998,
      "learning_rate": 5.482167869841611e-06,
      "loss": 3.1275,
      "step": 92100
    },
    {
      "epoch": 9.918200408997954,
      "grad_norm": 0.8838419914245605,
      "learning_rate": 5.1589268397801965e-06,
      "loss": 3.1385,
      "step": 92150
    },
    {
      "epoch": 9.923581961037563,
      "grad_norm": 0.8708315491676331,
      "learning_rate": 4.83568580971878e-06,
      "loss": 3.1325,
      "step": 92200
    },
    {
      "epoch": 9.928963513077171,
      "grad_norm": 0.9714722633361816,
      "learning_rate": 4.512444779657364e-06,
      "loss": 3.1291,
      "step": 92250
    },
    {
      "epoch": 9.93434506511678,
      "grad_norm": 0.8961527943611145,
      "learning_rate": 4.189203749595948e-06,
      "loss": 3.1339,
      "step": 92300
    },
    {
      "epoch": 9.939726617156388,
      "grad_norm": 0.9741976261138916,
      "learning_rate": 3.865962719534533e-06,
      "loss": 3.1428,
      "step": 92350
    },
    {
      "epoch": 9.945108169195997,
      "grad_norm": 0.8602545261383057,
      "learning_rate": 3.542721689473117e-06,
      "loss": 3.1453,
      "step": 92400
    },
    {
      "epoch": 9.950489721235604,
      "grad_norm": 0.8766176700592041,
      "learning_rate": 3.2194806594117013e-06,
      "loss": 3.1286,
      "step": 92450
    },
    {
      "epoch": 9.955871273275212,
      "grad_norm": 0.9090676307678223,
      "learning_rate": 2.8962396293502854e-06,
      "loss": 3.1159,
      "step": 92500
    },
    {
      "epoch": 9.961252825314821,
      "grad_norm": 0.8845322132110596,
      "learning_rate": 2.5729985992888694e-06,
      "loss": 3.1353,
      "step": 92550
    },
    {
      "epoch": 9.96663437735443,
      "grad_norm": 0.924277663230896,
      "learning_rate": 2.249757569227454e-06,
      "loss": 3.1229,
      "step": 92600
    },
    {
      "epoch": 9.972015929394038,
      "grad_norm": 0.8867099285125732,
      "learning_rate": 1.926516539166038e-06,
      "loss": 3.1248,
      "step": 92650
    },
    {
      "epoch": 9.977397481433645,
      "grad_norm": 0.9389845132827759,
      "learning_rate": 1.603275509104622e-06,
      "loss": 3.1412,
      "step": 92700
    },
    {
      "epoch": 9.982779033473253,
      "grad_norm": 0.9016231894493103,
      "learning_rate": 1.2800344790432064e-06,
      "loss": 3.1243,
      "step": 92750
    },
    {
      "epoch": 9.988160585512862,
      "grad_norm": 0.8891977071762085,
      "learning_rate": 9.567934489817906e-07,
      "loss": 3.112,
      "step": 92800
    },
    {
      "epoch": 9.99354213755247,
      "grad_norm": 0.896614134311676,
      "learning_rate": 6.335524189203748e-07,
      "loss": 3.126,
      "step": 92850
    },
    {
      "epoch": 9.998923689592079,
      "grad_norm": 0.9039609432220459,
      "learning_rate": 3.103113888589591e-07,
      "loss": 3.1234,
      "step": 92900
    },
    {
      "epoch": 10.0,
      "step": 92910,
      "total_flos": 7.76821211136e+17,
      "train_loss": 3.4538463850168144,
      "train_runtime": 79850.848,
      "train_samples_per_second": 37.232,
      "train_steps_per_second": 1.164
    }
  ],
  "logging_steps": 50,
  "max_steps": 92910,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 10000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 7.76821211136e+17,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}